Solar Live Steam Generation and Solar Bagasse Drying for South African Sugar Mills

Solar Live Steam Generation and Solar

Bagasse Drying for South African Sugar Mills

by

Willem Krog

March 2018

Thesis presented in partial fulfilment of the requirements for the degree

of Master of Engineering (Mechanical) in the Faculty of Engineering at

Stellenbosch University

Supervisor: Dr J.E. Hoffmann

Co-supervisor: Dr S. Hess

i

Declaration

By submitting this thesis electronically, I declare that the entirety of the work

contained therein is my own, original work, that I am the sole author thereof (save

to the extent explicitly otherwise stated), that reproduction and publication thereof

by Stellenbosch University will not infringe any third party rights and that I have

not previously in its entirety or in part submitted it for obtaining any qualification.

Date: ........March 2018.......................................

Copyright © 2018 Stellenbosch University

All rights reserved

Stellenbosch University https://scholar.sun.ac.za

ii

Abstract

Two solar thermal integration concepts have been identified as promising options

for implementation in South African sugar mills as a result of work within the

Sugarcane Technology Enabling Programme for Bio-Energy. The two integration

options are the drying of bagasse using solar heated air and the generation of live

steam using concentrating solar thermal collectors. This study further develops

and evaluates the two integration options.

Solar integration into the bagasse drying concept will help to save exhaust steam,

which can be used to dry bagasse. An evacuated tube air collector field was

simulated to assess the impact it could make on the bagasse drying system. It was

calculated that 3 140 ton of bagasse or 1020 ton coal can be saved through solar

thermal integration. If only exhaust steam is used in the bagasse drying system,

bagasse usage can be reduced with 5 %, but if the solar system is integrated it can

be decreased with 7.05 %.

The System Advisor Model was used to simulate two parabolic trough fields for

the live steam generation integration point, one for a normal sugar mill and one

for a mill with a back-end refinery. The simulations showed that the solar systems

have low capacity factors, ranging from 13 - 14.9 % depending on the mill and

time of operation. This was due to the low amount of direct normal irradiance

received in Durban, which severely hampers the performance of the solar system.

Three different solar live steam configurations were evaluated for the two mills,

each of which can save bagasse or coal and/or generate extra electricity. By using

the simulation results, it was determined that Configuration 1 can save 2 459.7 ton

coal for a normal mill and 3 248 ton coal for a mill with a refinery.

Configuration 2 can save 2 241 ton of coal and increase electricity exports with

257 % for a normal mill. For a mill with a refinery 3 072 ton coal can be saved

and electricity exports can be increased by 102 %. Configuration 3 can generate

the most extra electricity and would enable the mill to increase its electricity

exports with 297 % for the normal mill and 111 % for the mill with the refinery.

The simulation results were also used in an economic assessment for both of the

integration options. The assessment determined that none of the integration

options are financially feasible under current conditions. As none of the

integration points could achieve a levelised cost of heat lower than that of coal

(4.03 Euro-ct/kWh) or an internal rate of return higher than 10 %.

There is, however, a possibility that the integration points can become more

financially rewarding in the future, as the cost of solar thermal technology is set to

reduce significantly over the next 10 years. Furthermore, the possible carbon tax

which is to be implemented in South Africa will increase the cost of using coal,

making solar energy the cheaper option to supply thermal power.


iii

Opsomming

Twee son termiese integrasi konsepte is as belowende opsies vir implementering

in Suid-Afrikaanse suikermeulens deur ‘n vorige studie van die Suikerriet

Tengologie Instaatstellings Program vir Bio-Energie geïdentifiseer. Die twee

integrasie konsepte is die uitdroog van bagasse deur son verhitte lug en die

opwekking van hoë drukstoom deur gebruik temaak van gekonsintreerde

sonkrag. Hierdie studie ontwikkel die integrasie opsies verder en evalueer die

impak wat dit kan maak op ‘n suikermeule.

Son termiese energie kan die gebruik van uitlaatstoom in die verdrogings sisteem

verminder. ‘n Veld van vakuumbuis sonkollektors is gesimuleer om die potensiële

energie opbrengs wat dit kan bied te bepaal, asook die impak wat dit kan maak op

die verdrogingssisteem. Dit was bereken dat 3 140.62 ton bagasse of 1020.38 ton

steenkool gespaar kan word deur die sonkragsisteem. As net uitlaatstoom gebruik

sou word vir die verdrogings proses dan sal die gebruik van bagasse met 5 % kan

afneem, maar as son termiese energie ook gebruik word dan kan dit met 7.05 %

verminder.

Die System Advisor Model is gebruikom twee parabolise trog sisteme te simuleer

vir die hoë druk stoom opwekking stelsel, een sisteem vir ‘n gewone suiker meule

en een vir ‘n meule met ‘n suikerraffinadery. Die simulasies wys dat die

sonkragsisteme lae jaarlikse kapasiteitsfaktore het wat wissel tussen 13 – 14.9 %.

Die lae kapasiteitsfaktore in die simulasies is veroorsaak deur die lae jaarlikse

direkte normale sonsbestraling in Durban, wat die sonkragsisteme se prestasies

ernstig benadeel.

Drie verskillende konfigurasies vir die hoë druk stoom opwekking stelsel is

evalueer, die konfigurasies kan bagasse of steennkool besparaar en/of ekstra

elektrisiteit opwek. Die simulasie resulate was gebruik om te bereken dat

Konfigurasie 1, 2 459.7 ton steenkool kan spaar vir die gewone suikermeule en

3 248.5 ton steenkool vir die meule met die raffinadery. Konfigurasie 2 kan

2 241 ton steenkool spaar en die elektrisiteits uitvoere met 257 % vermeerder vir

die gewonemeule, terwyl dit 3072 ton steenkool kan spaar en die elektrisiteits

uitvoere met 102 % kan vermeerder vir ‘n meule met ‘n raffinadery.

Konfigurasie 3 kan die meeste ekstra elektrisiteit opwek, dit kan lei tot ‘n 297 %

toename in elektrisiteit suitvoere vir ‘n normale meule en ‘n toename van 111 %

vir ‘n meule met ‘n raffinadery.

Die ekonomiese evaluering toon dat nie een van die integrasiekonsepte finasieël

lonend sal wees onder huidgi ekondisies nie. Dit is as gevolg van die feit dat nie

een van die twee ‘n laer gebalanseerde koste van hitte kan bied as steenkool nie,

wat staan op 4.03 Euro-ct/kWh. Die integrasiekonsepte kan ook nie ‘n interne

oprengskoers hoër as 10 % behaal nie.


iv

Daar is egter hoop dat die integrasiekonsepte meer finansieel lonend sa lwees in

die toekoms as gevolg van die potensiële koste vermindering van sonkragsisteme

in die volgende 10 jaar. Suid-Afrika beplan ook om ‘n koolstofbelasting in te stel,

wat die gebruik van steenkool beboet en sodoende son termiese energie die

goedkoper opsie kan maak.


v

Acknowledgements

I would like to express my gratitude to my supervisor, Dr Jaap Hoffmann, and co-

supervisor, Dr Stefan Hess, for their support and guidance throughout this project.

Thank you for the opportunity to learn from you and all your help. I would also

like to thank Dr Katherine Foxon (SMRI) and Dr RynoLaubscher (Stellenbosch

University) for all their help and input regarding information and knowledge of

South African sugar mills.

I am grateful to the Centre of Renewable and Sustainable Energy Studies

(CRSES) and the Sugarcane Technology Enabling Programme for Bio-Energy

(STEP-Bio), which is co-funded by the Department of Science and Technology

(DST) and the South African sugar industry under the DST’s Sector Innovation

Fund for their financial support throughout this project.

To the team at the Solar Thermal Energy Research Group (STERG) – Thank you

for your friendship and informative talks over the last two years. I feel fortunate to

have been surrounded by people with a shared enthusiasm for renewable and

sustainable energy.

Lastly, I would like to thank my family and friends for their support and kindness

which carried me throughout this project. I would like to say a special thank you

to Maria for her endless love and emotional support.


vi

Table of Contents

Declaration .............................................................................................................. i

Abstract ................................................................................................................... ii

Opsomming ............................................................................................................ iii

Acknowledgements ................................................................................................ v

Table of Contents .................................................................................................. vi

List of Figures ...................................................................................................... viii

List of Tables ......................................................................................................... xi

Nomenclature ....................................................................................................... xii

1 Introduction ........................................................................................................ 1

1.1 Background ................................................................................................ 1

1.2 Motivation .................................................................................................. 2

1.3 Objectives .................................................................................................. 2

1.4 Methodology .............................................................................................. 3

1.5 Research Limitations ................................................................................. 4

2 Project Background ........................................................................................... 5

2.1 Sugar Milling Industry in South Africa ..................................................... 5

2.2 Generic Sugar Milling Process .................................................................. 6

2.3 Possible Solar Thermal Integration Points ................................................ 9

2.3.1 Live steam generation ................................................................... 9 2.3.2 Exhaust steam generation .............................................................. 9

2.3.3 Pre-heating of boiler feed water .................................................. 10 2.3.4 Drying of bagasse ........................................................................ 10 2.3.5 Drying of raw sugar ..................................................................... 11

2.3.6 Heating of clear juices ................................................................. 11 2.3.7 Integration points selected for further investigation ................... 11

3 Literature Study ............................................................................................... 13

3.1 Solar Energy ............................................................................................ 13

3.1.1 Solar technology review .............................................................. 13 3.1.2 Solar resource review .................................................................. 20

3.2 Bagasse Drying ........................................................................................ 24 3.2.1 Advantages of bagasse drying ..................................................... 24 3.2.2 Potential problems and disadvantages of bagasse drying ........... 25

3.2.3 Drying process ............................................................................. 28 3.2.4 Dryer types .................................................................................. 30

3.3 Cogeneration and Hybridisation with CSP .............................................. 33

4 Solar Bagasse Drying ....................................................................................... 38


vii

4.1 Integration Point ...................................................................................... 38

4.2 Drying Model ........................................................................................... 40 4.2.1 Simprosys .................................................................................... 40 4.2.2 Drying simulation setup with Simprosys .................................... 42

4.2.3 Simulation results ........................................................................ 43

4.3 Solar Field Modelling and Simulation ..................................................... 44 4.3.1 Modelling of an evacuated tube air collector .............................. 44 4.3.2 Matlab simulation results ............................................................ 48 4.3.3 System planning using Aircow ................................................... 50

4.3.4 Aircow results ............................................................................. 53

4.4 Solar Field Simulation ............................................................................. 56

4.4.1 Simulation setup with Solgain ..................................................... 56 4.4.2 Simulation results ........................................................................ 59

4.5 Effect on Sugar Mill ................................................................................ 61

5 Solar Live Steam Generation .......................................................................... 63

5.1 Integration Point ...................................................................................... 63

5.2 Increasing Electricity Production ............................................................ 65

5.3 System Advisor Model Simulation Setup ................................................ 67

5.4 System Advisor Model Simulation Results ............................................. 69

5.5 Effect on Sugar Mill ................................................................................ 72

6 Economic Assessment ...................................................................................... 75

6.1 Investment Costs ...................................................................................... 75 6.1.1 Specific investment costs of the bagasse drying solar field ........ 75 6.1.2 Specific investment costs of the live steam generation solar

field .............................................................................................. 76

6.2 Levelised Cost of Heat ............................................................................. 76

6.3 Internal Rate of Return ............................................................................ 79

7 Conclusion ........................................................................................................ 81

7.1 Summary of Findings .............................................................................. 81

7.2 Concluding Remarks ............................................................................... 82

7.3 Recommendations for Further Work ....................................................... 83

Appendix A: Airwasol Brochure ........................................................................ 84

Appendix B: Matlab Model Calculations .......................................................... 87

Appendix C: Matlab Model Comparison .......................................................... 91

Appendix D: Aircow Optimisation ..................................................................... 93

Appendix E: Comparison of Tracking Systems ................................................ 97

References ............................................................................................................. 98


viii

List of Figures

Figure 1: Modelling and simulating the solar bagasse drying integration option. .. 3

Figure 2: Locations of Sugar Mills in South Africa (Beukes et al., 2015) .............. 5

Figure 3: Simplified sugar milling processes (Hess et al. 2016) ............................. 6

Figure 4: BRTEM sugar mill steam network adapted from Hess et al. (2016) ....... 7

Figure 5: Schematic of a flat plate collector (Solar Advice, 2016). ...................... 14

Figure 6: Energy input and losses for an evacuated tube. ..................................... 15

Figure 7: (a) Schematic of a parabolic trough collector (Cabrera et al., 2013).

(b) Actual parabolic trough from Sunray Energy facility in Daggett

(Sun & Wind Energy, 2017). .............................................................. 16

Figure 8: (a) Reflection of sunrays onto absorber in linear Fresnel collector

(Electromagnetic Foundations of Solar Radiation Collection,

2017). (b) Linear Fresnel collector at Kimberlina, U.S.A. (CSP

World Organisation, 2015) ................................................................. 18

Figure 9: Central receiver system at Crescent Dunes (Solar Reserve, 2017). ....... 20

Figure 10: (a) Direct normal irradiance. (b) Global tilted irradiance. (Meyer,

2016) ................................................................................................... 21

Figure 11: KwaZulu Natal Solar Resource (GeoSUN, 2012). .............................. 22

Figure 12: DNI comparison of Upington and Durban (Meyer, 2016) ................... 23

Figure 13: (a) Annual GHI for Durban from 1994 - 2016. (b) Monthly average

GHI for Durban (Solargis, 2017) ........................................................ 23

Figure 14: (a) Bagasse to steam ratio as a function of the bagasse moisture

content. (b) Boiler fan power as a function of the bagasse moisture

content (Magasiner, 1987). ................................................................ 25

Figure 15: Simulation of different bagasse moisture contents for a boiler,

similar to the one assumed for the BRTEM model (Laubscher,

2017) ................................................................................................... 26

Figure 16: Drying rate under constant drying conditions (Tawfik et al., 2003) .... 28

Figure 17: The heating of bagasse under an oxygen atmosphere from Sosa-

Arnoa and Nebra, 2009) ...................................................................... 29

Figure 18: Co-current rotary dryer system ............................................................ 31

Figure 19: Pneumatic dryer system ....................................................................... 32


ix

Figure 20: Layout of the Borges Termosolar hybrid power plant (Biomass

Knowledge Portal, 2015). ................................................................... 33

Figure 21: Electricity output of the hybrid concept designed by (Peterseim et

al, 2014) .............................................................................................. 34

Figure 22: Simplified layout of CSP integration into a Brazilian sugar mill

(Burin, et al., 2016) ............................................................................. 35

Figure 23: LCOE and additional electricity produced for different solar

multiples (Burin, et al., 2016) ............................................................. 36

Figure 24: Basic schematic of solar bagasse drying integration point. ................. 39

Figure 25: Simprosys model layout. ...................................................................... 42

Figure 26: (a) Airwasol air collector. (b) One of the evacuated tubes in the

Airwasol air collector (Siems, 2017). ................................................. 44

Figure 27: Schematic of heat transfer resistances in an evacuated tube. ............... 45

Figure 28: Matlab simulation results for the evacuated tube ................................. 48

Figure 29: Extract of Paradis et al. (2015) experimental results. (a) Ambient air

temperature, simulation output temperature and experimental

output temperature as a function of time. (b) Tilted irradiance,

horizontal irradiance and reflected irradiance. (c) Wind speed

during testing. (d) Volume flow rate through evacuated tube. ........... 49

Figure 30: (a) Flow in a manifold’s lateral from Bajura& Jones (1976). (b)

Airwasol collectors connected in a row, with headers on the far left

and far right and supports in between (Siems, 2017). ......................... 52

Figure 31: Solgain system layout (Ilchmann et al., 2016). .................................... 58

Figure 32: Comparison of available solar radiation and thermal energy

delivered to process. ............................................................................ 60

Figure 33: Simulated solar thermal energy delivered to the drying system. ......... 61

Figure 34: Solar live steam integration into a generic sugar mill, adapted from

Hess et al. (2017). ............................................................................... 64

Figure 35: Schematic of Configuration 2 .............................................................. 65

Figure 36: Schematic of Configuration 3 .............................................................. 66

Figure 37: Simulated annual solar field output ...................................................... 70

Figure 38: Simulated annual solar field output for an east-west tracking axis

system ................................................................................................. 71

Figure 39: Possible electricity exports. .................................................................. 74


x

Figure 40: The transmissivity of Borofloat 33 over various wavelengths

(Schott, 2017). ..................................................................................... 88

Figure 41: Change in collector outlet temperature as flow rate changes as

predicted by model of Paradis et al. (2015) ........................................ 91

Figure 42: Change in collector outlet temperature as flow rate changes as

predicted by model used in thesis. ...................................................... 91

Figure 43: Change in collector efficiency as flow rate changes as predicted by

model of Paradis et al. (2015) ............................................................. 92

Figure 44: Change in collector efficiency as flow rate changes as predicted by

model used in thesis. ........................................................................... 92

Figure 45: Aircow’s Global system considered for mass flow optimisation as

shown in the Aircow manual (Fraunhofer ISE, 2017). ....................... 93

Figure 46: Comparison of different number of collectors in a row with

optimised flow as shown in the Aircow manual (Fraunhofer ISE,

2017). .................................................................................................. 96

Figure 47: Comparison of the performance of the two tracking systems on 21

June. .................................................................................................... 97

Figure 48: Comparison of the performance of the two tracking systems on 20

December. ........................................................................................... 97


xi

List of Tables

Table 1: Sugar milling parameters used for this study. ........................................... 9

Table 2: Typical design and performance values for rotary dryers (Bruce &

Sinclair, 1996) ..................................................................................... 31

Table 3: Typical design and performance values for a pneumatic dryer (Bruce

& Sinclair, 1996) ................................................................................. 32

Table 4: Input values for Simprosys simulation. ................................................... 43

Table 5: Simprosys outputs.................................................................................... 43

Table 6: Airwasol evacuated tube characteristics .................................................. 45

Table 7: Data points for Aircow input ................................................................... 53

Table 8: Evacuated tube quadratic model coefficients as determined by

Aircow. ................................................................................................ 54

Table 9: Converted coefficients for liquid collector simulation. ........................... 56

Table 10: Inputs for Solgain .................................................................................. 59

Table 11: Simulation characteristics as calculated by Solgain .............................. 60

Table 12: Solar field parameters for the two SAM simulations. ........................... 69

Table 13: SAM simulation results ......................................................................... 69

Table 14: SAM simulation results for an east-west tracking system. .................... 71

Table 15: Impact of solar live steam generation. ................................................... 72

Table 16: Values used to determine electricity generation. ................................... 73

Table 17: LCOH of the solar thermal integration options. .................................... 78

Table 18: IRR for the integration options. ............................................................. 79


xii

Nomenclature

Symbol Description Unit

A Cross sectional area m2

Aa Outside area of the absorber m2

cad Specific heat capacity of dry air J/(kg K)

cam Specific heat capacity of the moisture in air J/(kg K)

cbd Specific heat capacity of dry bagasse J/(kg K)

cbm Specific heat capacity of the moisture in bagasse J/(kg K)

ceff Effective collector heat capacity kJ/(m2 K)

cp,HTF Specific heat capacity of the heat transfer fluid J/(kg K)

cp,c Specific heat capacity air in collector J/(kg K)

C1 Linear heat loss coefficient W/(m2 K)

C2 Quadratic heat loss coefficient W/(m2 K2)

Cm Mass flow dependant heat loss coefficient h/kg

Cn Capital investment costs Euro

d Nominal discount rate %

Dg Glass tube outer diameter m

Eb Energy emitted for specific wave length W

f Friction factor -

F Solar fraction -

G Global horizontal irradiance W

Gb Direct horizontal irradiance W

Gd Diffuse horizontal irradiance W

Gbn Direct normal irradiance W

Gbt Direct irradiance on tilted collector W

Grt Reflected irradiance on tilted collector W

Gst Diffuse irradiance on tilted collector W

Gt Solar irradiance on tilted collector W

hcond Enthalpy of condensate kJ/kg

hexhs Enthalpy of exhaust steam kJ/kg

hls Enthalpy of live steam kJ/kg


xiii

hBFW Enthalpy of boiler feed water kJ/kg

hrad Heat transfer coefficient between glass and sky W/(m2 K)

∆ℎ𝐴 Latent heat of vaporisation kJ/kg

∆ℎ𝐵 Heat of sorption kJ/kg

𝐼𝐴 Enthalpy of drying air kJ/kg

𝐼𝐵 Enthalpy of bagasse kJ/kg

Is Irradiance through glass tube W

k Thermal conductivity W/(m K)

Kb Direct irradiance incidence angle modifier -

Kd Diffuse irradiance incidence angle modifier -

�̇� Air mass flow in collector kg/s

�̇�𝑎 Air mass flow in dryer kg/h

�̇�𝑏 Bagasse mass flow in dryer kg/h

�̇�𝑏𝑎𝑔 Bagasse mass flow into boiler ton/h

�̇�𝑒𝑥ℎ𝑠 Exhaust steam flow rate ton/h

�̇�𝑙𝑠 Live steam flow rate ton/h

�̇�𝐵𝐹𝑊 Boiler feed water flow rate ton/h

�̇�𝐻𝑇𝐹 Mass flow in parabolic trough loop kg/s

Mev Moisture evaporated kg/h

n Year -

N Project financial life span years

NL Number of loops -

p Pressure Pa

∆𝑝𝑐𝑜𝑙𝑙 Pressure drop in collector Pa

∆𝑝𝑠𝑦𝑠 System pressure drop Pa

Pfan Fan power W

Qag Radiation heat transfer between absorber and glass W

Qc Indirect supplied heat to dryer kW

Qcoll Thermal energy delivered by the solar field kW

Qgain Collector energy gain W

Ql Heat losses in dryer kW

Qm Mechanical input into dryer kW


xiv

Qn Annual yield of the parabolic trough system kWh

Qneed Thermal energy needed by the process kW

Qp Thermal energy delivered to the process kW

QPT Thermal power of parabolic trough field kW

Qrad Radiation onto absorber W

Qrad,a Radiation absorbed by glass W

Qsun Sun’s energy delivered to solar field kW

Qt Heat carried into dryer by transport device kW

Qth,dry Thermal energy for dryer kW

ra Outer radius of absorber m

rg Inner radius of glass tube m

R1 Linear pressure drop coefficient Pa/(kg/h)

R2 Quadratic pressure drop coefficient Pa/(kg2/h2)

Rag Resistance to heat transfer from absorber to glass K/W

Rair Resistance to heat transfer from absorber to air K/W

Rgsky Resistance to heat transfer from glass to sky K/W

Rg∞ Resistance to heat transfer from glass to ambient air K/W

Rsys System flow resistance Pa/(m3/h)2

T Temperature K

Ta Absorber temperature K

Tair Air temperature in collector K

Tb Bulk temperature K

Tf Mean collector temperature K

Tg Glass temperature K

Tsky Sky temperature K

T∞ Ambient temperature K

v Air velocity in collector tube m/s

vg Gas velocity m/s

X Moisture content of bagasse %

Y Moisture content of air %


xv

Greek

Symbol Description Unit

𝛼 Absorptance -

𝛽 Volume expansivity 1/K

𝛽c Collector tilt -

𝜀𝑎 Emissivity of absorber -

𝜀𝑔 Emissivity of glass -

𝜂𝑁𝐶𝑉 Boiler efficiency based on net calorific value of fuel %

𝜂 Collector efficiency %

𝜂0 Conversion factor %

𝜂𝑠 Annual system efficiency %

𝜂𝑢 Degree of utilisation %

𝜃 Incidence angle °

𝜃𝑠 Zenith angle of the sun °

𝜆 Wavelength 𝜇𝑚

𝜈 Kinematic viscosity m2/s

𝜌 Reflectivity -

𝜌𝑔𝑟𝑑 Ground reflectivity -

𝜏 Transmissivity -

Dimensionless Numbers

Nu Nusselt number

Pr Prandtl number

Ra Raleigh number

Re Reynolds number

Abbreviations

AE Additional Electricity

BPST Back pressure steam turbine

CSP Concentrated solar power

CEST Condensing extract steam turbine

DNI Direct normal irradiance


xvi

ECO Economiser

EfB Energy from Boiler

EfW Energy from Waste

EVAP Evaporation

GCV Gross calorific value

GHI Global Horizontal irradiance

GTI Global tilted irradiance

IPP Independent power producer

IPPPP Independent power producer procurement plan

IRR Internal rate of return

LCOE Levelised cost of electricity

LCOH Levelised cost of heat

NCV Net calorific value

REIPPPP Renewable Energy Independent Power Producer Procurement Program

SAM System advisor model

SH Super Heater

SMRI Sugar Milling Research Institute

STEP-Bio Sugarcane Technology Enabling Programme for Bio-Energy

ST Solar Thermal

TMY Typical meteorological year


1

1 Introduction

1.1 Background

The South African sugar industry wants to be more cost competitive and energy

efficient. To support this endeavour, the South African Department of Science and

Technology fund the Sugarcane Technology Enabling Programme for Bio-Energy

(STEP-Bio) with the Sugar Milling Research Institute (SMRI) coordinating the

research within the program. The overall aim is to increase the revenue per unit

sugarcane processed by developing fully integrated sugarcane bio-refineries. This

will also allow the sugar milling industry to diversify their income, which can

help with economic growth and lessen the impact of low sugar production due to

droughts, as in 2016 and 2017.

The aim of the research reported in this study is to assess to what extent solar

thermal energy can be used to reduce the running costs of South African sugar

mills and to open up additional income streams. This can ensure sustainability and

will make the industry less susceptible to South Africa’s increasing energy prices.

In initial studies on the topic by Beukes et al. (2015) and Hess et al. (2016), along

with feedback from the STEP-Bio steering committee, two solar integration

concepts have been identified as plausible options for implementation in the near

future. These are: the generation of live steam using concentrated solar power

(CSP) collector technology and the drying of bagasse using solar heated air.

Renewable energy has received a lot of focus the past few decades and the 2015

Paris Climate Conference put great emphasis on countries incorporating

renewable and sustainable technology into their energy sectors. There are various

renewable energy resources which can be used instead of fossil fuels, the

challenge, however, lies in creating economically viable solutions in an ever

growing and competitive environment. Various sources believe that renewable

energy will be able to meet most of the growing demand in energy, and this at

equal or lower prices than conventional energy sources (Kalogirou, 2009),

painting a bright future for renewable energy technologies.

The solar generation of live steam aims to increase electricity production in order

to sell it to the grid. This will allow the mills to expand their income stream. The

drying of bagasse will increase the calorific value of the bagasse, which can lead

to bagasse or coal savings and higher boiler efficiencies.

This project builds on the work of Beukes et al. (2015) and Hess et al. (2016) and

will focus on the above-mentioned solar thermal integration points and their effect

on the sugar milling process. This study worked closely with the SMRI regarding

the integration of the solar thermal technology, to ensure that the solar integration

options can provide realistic solutions and that it suits the sugar mills’ operating

conditions.


2

1.2 Motivation

The Department of Energy created a programme for independent power producers

(IPP) to help increase South Africa’s electricity capacity. The programme, called

the Independent Power Producer Procurement Programme (IPPPP), requests bids

for cogeneration projects. The programme caters for a maximum of 800 MWel in

the first bidding round and this was later increased to 1800 MWel. Approximately

25% of this capacity is allocated for combined heat and power projects. Priority is

given to projects where the energy output can be increased by upgrading existing

equipment and improving operating efficiencies. The STEP-Bio programme

seems to fit all the criteria of this program. This would enable sugar mills to

export electricity to the grid and open up a new income stream.

Large scale Solar Process Heat (SPH) integration within South Africa’s industrial

sector is still in a developmental phase, with no recorded high temperature

processes using solar energy.This project can help the country realise its potential

in terms of SPH integration. It can put the sugar milling industry in the spotlight

as an innovative and green industrial sector and serve as a demonstration of how

solar energy can be utilised for process heat.

1.3 Objectives

This study focuses on solar bagasse drying and solar live steam generation in

order to give a sugar factory a feasible solution for increasing their revenue per

unit sugarcane processed. The objectives of this study are to:

• Further develop specifically adapted solar thermal system concepts for the

two most promising integration points.

• Create detailed simulation results for the annual heat production of the

solar thermal systems in the integration concepts.

• Determine the impact the solar process heat technologies can have on the

sugar mill.

• Determine the financial feasibility of the two integration points.

The above objectives will allow this study to paint a clear picture of how solar

thermal energy can be incorporated in South African sugar mills, the impact it can

have and if it is financially viable.


3

1.4 Methodology

The first step of this study was to review the work of Beukes et al. (2015) and

Hess et al. (2016) to better understand the potential of solar thermal integration

options for South African sugar mills. A brief study on the South African sugar

industry was done to be able to understand and work with the generic sugar mill

model of Starzak and Zizhou (2015). In order to further develop the two chosen

solar thermal integration points a good understanding of solar energy, bagasse

drying and cogeneration with CSP was needed, and therefore researched.

To further develop the solar bagasse drying integration point a drying model of

the bagasse dryer is created in Simprosys(Simprotek, 2006). By simulating this

model the needed thermal energy and air flow to dry the bagasse can be

determined. Figure 1, shows how this was used along with other models to setup a

simulation in Solgain (Ilchmann et al., 2016).

Figure 1: Modelling and simulating the solar bagasse drying integration option.

The thermal power delivered by the solar field is used to determine how much

energy the integration point can save for the bagasse drying system. Furthermore,

the amount of bagasse and coal that can be saved by the drying system is

calculated.

To further develop the solar live steam generation integration point two new

turbine configurations and their possible impact on the sugar mill are discussed.

System Advisor Model (SAM) (National Renewable Energy Laboratory , 2017) is

used to simulate the suggested solar field. The simulation gives the thermal power

generated by the solar system as one of its outputs, and this is used to determine

the effect the integration point can have on the sugar mill.

The simulation results of the two integration points are then used to determine the

financial feasibility of the two integration points. Literature sources and solar

project databases are assessed in order to try and get an as accurate as possible

estimate of the investment costs. An economic assessment is done by determining


4

the levelised cost of heat (LCOH) and internal rate of return (IRR) for both

options.

Finally, conclusions are drawn regarding the technical and financial feasibility of

solar thermal integration into South African sugar mills.

1.5 Research Limitations

This study is based on a generic South African sugar mill as described by Starzak

and Zizhou (2015), specifications regarding this generic mill will be discussed in

section 2.1. In reality no two sugar mills are alike, varying in size and technology.

This is why a mill which represents the average industry practices was chosen.

Therefore, the integration options specified in this study will have to be adapted

for a specific mill, if it is considered implementing the proposed solar thermal

technology. This will mean having to re-evaluate certain assumptions and the

solar thermal technology’s suitability for the mill. Furthermore, the solar resource

at the mill’s location will have to be evaluated and the simulations and economic

assessment will have to be done again, incorporating all the new factors.


5

2 Project Background

This section will give more information regarding the sugar milling industry, the

various processes in the sugar mills and the work that Beukes et al. (2015) and

Hess et al. (2016) did for the STEP-Bio project.

2.1 Sugar Milling Industry in South Africa

The agriculture and agro-processing of sugarcane in South Africa are very

important, employing almost 80 000 people and making important contributions

to the national economy and especially the provincial economies of KwaZulu

Natal and Mpumalanga (Smith et al., 2016). South Africa has 14 sugar mills

scattered across the two above mentioned provinces, as can be seen in

Figure 2.

Figure 2: Locations of Sugar Mills in South Africa (Beukes et al., 2015)

In 2016 a total of 14.86 million tons of sugarcane were processed by these mills,

producing 1.64 million raw and refined sugar as well as a wide range of sugar by-

products (Smith et al., 2016). However, 2016 and 2017 were both below average

years in terms of sugar production, due to the severe drought in the area during

these two years (Singels et al. , 2017). The industry actually has the capacity to

process 22 million tons of sugarcane per crushing season and prospects for future

production looks good due to the end of the drought (Madho et al., 2017).

Each mill’s capacity differs, ranging from 90 to 550 tons of cane processed per

hour (Smithers, 2014). In the 2016 season the average sugar mill ran for 233 days


6

Cane

Draft juice Clear juice Syrup Raw sugarRaw

sugar

Bagasse

Condensate

Livesteam Exhaust steam

Vapour Molasses

Electricity

Make-upwater

Coal

CrystallizationClarification

Steamboiler

Preparation

Electricitygeneration

Extraction DryingEvaporation

and had an overall time efficiency (OTE) of 80.83 % (Smith et al., 2016). The

crushing season, during which the sugarcane is harvested and processed by the

mills, stretch between March and December and during this time the sugar mills

run day and night. Most South African cane producers use manual harvesting

techniques to get the sugarcane; this usually requires burning the sugar cane field

before hand, making it easier for the workers to harvest. The cane is then cut,

bundled and transported to the mills. There are, however, producers which use

motorised harvesters and harvests the cane while it is still green.

2.2 Generic Sugar Milling Process

As mentioned in Section 1.4, Starzak and Zizhou (2015) created a model which

represents a theoretical South African diffuser sugar mill as part of their study

called the Biorefinery Techno-Economic Modelling Project (BRTEM). The

process parameters used for this study are all results from the BRTEM model,

unless stated otherwise. It is important to note that the BRTEM model under

predicts the energy consumption of a real factory, as the modelled sugar milling

processes are more energy efficient than what they are in reality (Foxon, 2017).

The BRTEM model was designed and validated for describing the process flow

characteristics of a generic sugar mill, with less emphasis on validating the energy

consumption of the mill. Figure 3 shows the sugar milling processes for the

generic South African mill.

Figure 3: Simplified sugar milling processes (Hess et al. 2016)

During the cane preparation process dirt and rocks are removed from the

harvested cane, where after it is shredded in order to ease extraction. Sucrose is

leached from the shredded cane through the addition of imbibition water to form

the draft juice during the extraction process. After extraction the shredded cane

passes through roller mills to press out remaining sucrose and to reduce the cane’s

moisture content. The shredded cane, now called bagasse, and sucrose are then

split up, with the bagasse heading to the boiler house where it is used as fuel. The

draft juice goes on to be clarified in order to remove impurities. The clear juice

then enters an evaporation process where it forms a syrup. The syrup is then


7

Steamboiler

Primemovers

G

Exhaust steam

Condensate

Live steam 390 °C | 31 bar |96.76 t/h

121 °C | 2 bar

121 °C | 2 bar

Boilerfeed water

113 °C | 2bar96.96 t/h

Bagasse | 43.63 t/h

Coal

Make-upwater

25 °C | 2 bar6.1 t/h

Flue gas

Feed watertank

Clear juicepre-heater

1st Effectevaporator

Sugardrying

Let downvalve

Turboalternators

crystalized by boiling it under a vacuum, forming raw sugar and molasses (Rein,

2007). The raw sugar is separated from the molasses and the surface moisture

dried off.

There are two types of mills in use worldwide, namely: milling tandem systems

and diffuser systems. These systems differ in the way the sugars are extracted

from the sugarcane. The milling tandem system uses a series of mills which

mechanically squeezes the juice out of the sugarcane. The diffuser relies on the

diffusion and leaching of sugars to the imbibition water it soaked in (Oliverio et

al., 2014). Figure 3 represents a generic South African diffuser mill, because the

Extraction process relies on diffusion, the processes thereafter is to create raw

sugar from the imbibition water and sugarcane juice mixture.

It is important to distinguish between the two types of mills, because it has a

major influence on the mill’s energy and steam consumption. The milling tandem

sugar mills rely on live steam to drive the mechanical mills, while diffuser mills

uses lower temperature steam for heating the imbibition water (Foxon, 2017).

Figure 4 shows a schematic of the steam network of the generic South African

mill as specified in the BRTEM model. A sugar mill’s main energy source is

bagasse, the fibrous residue left of the sugarcane after the sugar extraction

process. Some of the mills export their bagasse, it can be sold to produce fertiliser,

animal feed, or paper, alternatively the mill can use it to create bio-ethanol. The

mills that do export their bagasse need extra fuel and in South Africa the sugar

mills use coal as an auxiliary fuel, because of its abundance and low price in the

country.

Figure 4: BRTEM sugar mill steam network adapted from Hess et al. (2016)


8

Live steam is used to drive three prime movers (cane knives, cane shredder and

dewatering mill) and a back pressure steam turbine (BPST). The three prime

movers prepare the sugarcane for sugar extraction while the turbine is connected

to an electric generator which supplies electricity to the mill. The exhaust steam

from the prime movers and turbine is used to supply thermal energy to the sugar

extraction and production processes via various heat exchangers. If the sugar

extraction processes need more exhaust steam than can be delivered by the prime

movers and turbine, live steam passes through a let-down station where it is

turned into exhaust steam.

The BRTEM model assumes that there is no need for let-down steam; however, in

reality most sugar mills are designed so that all the turbines do not supply more

than the lowest exhaust steam demand under relatively normal conditions. This

relates to 75 % - 95 % of the normal exhaust steam demand, therefore, the live

steam let-down valve is used to supply the remaining 5 % - 25 % of the necessary

exhaust steam (Rein, 2007; Foxon, 2017). For this study it is assumed that the let-

down valve needs to supply 10 % of the exhaust steam on average for the generic

sugar mill.

The system described above and by the BRTEM model is for a generic South

African sugar mill, however, four of the mills have back-end sugar refineries,

where the raw sugar produced by the mill is refined to white sugar. These sugar

mills tend to use a lot of extra coal, since they also operate outside of the crushing

season, when there is very little bagasse available. Refineries need exhaust steam

to provide thermal energy for the sugar refining process, as well as some

electricity. This means that sugar mills with back-end refineries need to produce

more steam and electricity during the crushing season, as can be seen in Table 1.

Outside of the crushing season the steam and electricity production is significantly

lower, since only the refinery is in operation as it processes the raw sugar

produced by other mills which do not have back-end refineries.

Table 1 shows the difference in energy needs between the normal mill and the

mill with a refinery. The values for the normal mill come from the BRTEM model

simulation results, while the refinery values were calculated for this study as

described below.

Rein (2007) gives the mass flow ratio between exhaust steam used in a refinery

and raw sugar refined as 1:1. The BRTEM model’s simulation results show that

the mill produces 29.08 ton/h raw sugar. Assuming that the refinery only has the

capacity to process the 29.08 ton/h raw sugar coming from the mill, its steam

consumption will be 29.08 ton/h. It is important to note that the sugar mill does

not operate outside of the crushing season, only the refinery does, as it refines the

raw sugar from other factories without back end refineries.


9

Table 1: Sugar milling parameters used for this study.

System

Live Steam Heat

Demand Steam Usage

Cane Crushed + Raw Sugar

Refined

Crushing

Season

[MW]

Outside

Crushing

Season

[MW]

Crushing

Season

[ton/h]

Outside

Crushing

Season

[ton/h]

Crushing

Season

[ton/h]

Outside

Crushing

season

[ton/h]

Sugar Mill 73 0 95 0 244.18 0

Sugar Mill

with

Refinery

95 22 125 29.08

244.18

+

29.08

0

+

29.08

2.3 Possible Solar Thermal Integration Points

As mentioned before, Beukes et al. (2015) and Hess et al. (2016) looked at

various options for solar thermal integration for South African sugar mills. They

came up with six different integration options, they are: live steam generation,

exhaust steam generation, pre-heating of boiler feed water, drying of bagasse,

drying of raw sugar and the heating of clear juices. This section will very shortly

explain each of the integration options as they developed it.

2.3.1 Live steam generation

For this integration point a concentrating solar system will heat up thermal oil

(used as heat transfer fluid) which will be used to produce steam in a kettle type

heat exchanger. The steam will have to be of a high enough pressure to pass

through the pressure valve and enter the conventional steam system.

By providing extra live steam through the solar thermal system, the boiler’s load

can be reduced. This will allow the mill to save either bagasse or coal. The mill

could produce extra electricity as well, if there was a process which could use the

extra exhaust steam from the BPST or if an alternative turbine configuration can

be installed, this will be discussed further in Section 4.

2.3.2 Exhaust steam generation

In this possible solar integration point stationary concentrating collectors or

stationary non-concentrating collectors can be used to supply the necessary

thermal energy. These technologies are considerably less complicated that what is

needed for the live steam generation point, because of the relatively low

temperature of the exhaust steam. The lower temperature and pressure of exhaust

steam allows for direct steam generation technology to be used, eliminating the

need for a kettle type heat exchanger and expensive heat transfer fluids.


10

This integration option is similar to the previous one, in the sense that it lessens

the boiler’s load, however, for this integration option it is to a lesser extent and the

solar thermal energy is provided as exhaust steam. By producing exhaust steam,

the mill can reduce the amount of live steam that needs to be let down and

therefore the amount of live steam that needs to be generated. The solar system

can provide the extra exhaust steam needed by the sugar extraction processes,

which the prime movers and turbine cannot provide. This will allow the mill to

save bagasse or coal.

The exhaust steam integration option could allow the mill to produce extra

electricity if there was an alternative turbine configuration which exhausted the

live steam in the turbine to below atmospheric conditions. This will then allow the

solar system to provide the necessary exhaust steam to the mill which the turbine

usually provided.

2.3.3 Pre-heating of boiler feed water

For this integration point concentrated solar collectors can be used to heat the

boiler feed water to 200 °C. This can be done by using pressurised water or

thermal oil as heat transfer fluid. The heat transfer fluid heats up the feed water

through a cost-effective plate heat exchanger, which has to be located downstream

from the feed water pump in order to avoid steam forming in the heat exchanger.

In the sugar milling industry the role of feed water pre-heating is usually fulfilled

by economisers, which recovers heat from the boiler’s flue gas to heat up the feed

water. As mentioned before, the BRTEM model does not take energy efficiency

measures like this into account, which opens the door for solar thermal

integration. However, it is good practice to implement energy efficiency measures

before solar thermal integration is considered.

2.3.4 Drying of bagasse

Bagasse moistures typically vary between 46 % and 52 % as they leave the

dewatering mills for the boiler. In the 2015/2016 crushing season the average

moisture content of the bagasse from all the mills was 51 % (Smith et al., 2016).

Bagasse can be dried to a technical lower limit of 30 % moisture content (Rein,

2007), but for the South African sugar milling industry the limit is set to 40 %.

This is the bottom range for which the boilers were designed and they might

struggle to perform properly using bagasse with lower moisture contents (Foxon,

2017). Bagasse drying is also not included in the BRTEM model, this is due to the

fact that it is not common practice in South Africa.

Bagasse drying has numerous advantages; it can improve boiler efficiency, reduce

fuel consumption, create higher flame temperatures and reduce excess air

requirements. By drying bagasse from 51 % to 40 % the bagasse gross calorific

value can increase by 24 % (Beukes et al., 2015). For the drying system a


11

stationary solar air collector can be used to heat air which would then pass

through a rotary dryer where the bagasse is dried.

2.3.5 Drying of raw sugar

Sugar drying is a very common in South African sugar mills and is therefore

included in the BRTEM model. Exhaust steam is used to heat air which then

enters a rotary dryer where the sugar is dried. Stationary solar air collectors can be

used to pre-heat the air, which would reduce exhaust steam consumption for the

sugar drying process. This integration option can work out very cheap for the

sugar mills due to the fact that dryers and exhaust steam air heaters are already

installed. Only a small solar field would be required for this integration point,

because of sugar dryings low thermal requirements.

2.3.6 Heating of clear juices

For his integration concept stationary solar collectors can be used to pre-heat the

clear juices to 110 °C. A heat transfer fluid, like pressurised water can be heated

up and pass through a heat exchanger to heat up the clear juices. This will happen

between the clarification stage and the evaporation stage. This will allow the mill

to save exhaust steam, as less thermal energy will be required to boil the heated

clear juices.

For this integration point it is important that the heat exchanger is bypassed during

times when the solar field cannot deliver the needed 110 °C. This would avoid a

heat flux from the process to the collector loop.

2.3.7 Integration points selected for further investigation

The STEP-Bio steering committee was very interested in the solar live steam

generation and solar bagasse drying integration points (Hess, 2016). They felt that

it warranted further investigation based on the results of the above mentioned

studies.

From Hess et al.’s (2016) economic assessment, the live steam generation option

faired relatively well considering the return on investment. However, it was still a

more expensive option to provide thermal energy compared to coal and requires

the highest capital investment. What interests the STEP-Bio steering committee is

the option of using solar live steam generation in a setup which will allow for

more electricity to be generated. This option was discussed in Hess et al. (2016)

and will be explained in detail in Section 4.

Bagasse drying faired very well in the economic assessment mentioned above. It

finished as second cheapest option and with the second highest return on

investment. Only the sugar drying option did better. Both sugar and bagasse


12

drying were considered for further development as they would make excellent

demonstration plants to showcase solar thermal technology’s benefits. In the end

the steering committee decided on the bagasse drying option as it offers multiple

new benefits for the sugar mills.


13

3 Literature Study

This section looks at information found in books, journal articles, reports and

conference papers regarding solar energy, bagasse drying and co-generation and

hybridisation with CSP.

3.1 Solar Energy

This subsection will contain information about various solar thermal technologies

commercially available and briefly explain how each technology works. The solar

resource available in South Africa will also be discussed, specifically looking at

KwaZulu Natal where most of the sugar mills are situated. It will explain the

difference between the types of irradiance measured and what is used by the

different technologies.

3.1.1 Solar technology review

Solar thermal systems convert sunlight into a heat source which can be used to

drive various thermal processes. Non-concentrating collectors like flat plate and

evacuated tube collectors are usually associated with low-medium temperature

applications and is mostly utilised for domestic uses and process heat applications.

Concentrating collectors like parabolic troughs, central receivers and linear

Fresnel systems are usually associated with high temperatures and are mostly

utilised for power generation. Beukes et al. (2015) and Hess et al. (2016)

suggested that stationary, non-tracking collectors be used for the bagasse drying

integration point and concentrating, tracking collectors be used for the solar live

steam generation. This section will give a brief overview of the above mentioned

collector technologies.

Flat Plate Collector:

A flat plate collector is arguably the simplest collector available. It basically

consists of a transparent cover, a dark absorber plate, flow tubes and insulation as

can be seen in Figure 5. It is due to this simplicity that it is one of the most

common collectors and is used mostly for domestic hot water systems for single

family homes (Joubert et al., 2016).

The transparent cover lets solar radiation pass through, allowing it to fall on the

absorber. The absorber heats up due to its high radiation absorptivity, a heat

transfer fluid passes through the flow tubes to transport the heat/thermal energy to

where it is needed. Unfortunately, not all the energy can be transferred to the heat

transfer fluid, there are various thermal and optical losses in the collector. Some

of the radiation is reflected and absorbed by the transparent cover and there are

convection, conduction and radiation losses from the absorber. The cover helps to

reduce convection losses by keeping air stagnant in the collector, it also reduces


14

radiation losses by being transparent to the short waves received from the sun, but

being nearly opaque to the long wave radiation emitted from the absorber

(Kalogirou, 2009). The insulation on the sides and bottom of the absorber helps to

decrease conduction losses.

Figure 5: Schematic of a flat plate collector (Solar Advice, 2016).

Most collectors have glass covers with low iron content (which have a higher

radiation transmittance than high iron content glass) however, there are low

temperature collectors which use plastic covers as well. To allow the absorber to

absorb as much of the solar radiation as possible and emit as little radiation as

possible, it has a selective coating. The selective coating along with the limited

radiation losses through the cover allows the collector to reach temperatures up

100 °C (Duffie and Beckman, 2006). This makes it suitable for various low

temperature applications, like cleaning, cooling and drying (Weiss and Rommel,

2008).

One of the main advantages of the flat plate collector is that it is inexpensive. It

can use water, water-glycol mixtures or air as heat transfer fluid, all of which are

very cheap. They are fixed collectors, so no expensive tracking systems are

required and they do not require much maintenance (Kalogirou, 2009).

Evacuated Tube Collector:

Evacuated tube collectors are a non-tracking technology, like flat plate collectors,

but can reach considerably higher temperatures (up to 200 °C) due to better heat

loss prevention measures. A drawback of flat plate collectors are that they perform

poorly in cold, windy and cloudy conditions due to high heat losses (Kalogirou,

2009). Evacuated tubes can perform better in these adverse conditions, due to the

fact that the absorber is enclosed in a glass tube under vacuum, minimizing

convection heat losses. The glass tube allows the sun’s rays to pass through it and

heat up the absorber and also reduces radiation losses as the transparent cover

does for the flat plate collector. Figure 6 shows the energy input and losses of the


15

evacuated tube. The absorber also has a selective coating with low emissivity,

which further reduces radiation losses.

Figure 6: Energy input and losses for an evacuated tube.

Evacuated tubes collectors have similar advantages as the flat plate collectors;

they are also relatively inexpensive and do not require much maintenance (Sabiha

et al., 2015). If one of the tubes in the collector breaks or need to be replaced, then

only that tube can be replaced, there is no need to replace the whole collector. The

glass is quite fragile though and can break due to hail or poor handling (Sabiha et

al., 2015).

Various heat transfer fluids can also be used for evacuated tube collectors, like

water and water-glycol mixtures, for higher temperature applications above

100 °C, air and pressurised water can be used (Joubert et al., 2016). Sabiha et al.

(2015) suggests that evacuated tube collectors should be used for processes with a

continuous and ample load which will ensure that the heat transfer fluid does not

become too hot. For lower temperature applications where water or a water-glycol

mixture is used it is essential to keep the temperature below 100 °C to prevent

over boiling. Boiling in the collector can expose weaknesses in the material and

damage the collector and vacuum.

Evacuated tube collectors can be used for the same applications as flat plate

collectors. They are especially useful if higher temperatures are needed or smaller

solar fields are required, because of their higher operational temperature (Weiss

and Rommel, 2008).

Parabolic Trough Collector:

Parabolic trough collectors are the most mature CSP technology in use today. In

2015, just more than 80 % of the installed and planned CSP capacity was in the


16

form of parabolic trough collectors (Liu, et al., 2016). Projects like the SEGS

plants, Nevada Solar One and Andasol helped to prove that parabolic troughs

using thermal oils are trustworthy technology for power generation. Parabolic

troughs are also used for medium and high temperature solar process heat

applications, with the database for applications of solar heat integration in

industrial processes (AEE Intec, 2017) listing 45 parabolic trough plants.

Parabolic troughs are classified as a line focusing technology, consisting out of a

parabolic mirror, receiver tube, support structure, tracking axis and supports as

can be seen in Figure 7 a and b. The parabolic mirrors or reflective coating

reflects sunlight onto a receiver tube to heat it up to temperature slightly above the

needed heat transfer fluid temperature. Temperatures above 500 °C can be

reached, with coatings which do not degrade at such high temperatures being

commercially available. However, it is the heat transfer fluid’s maximum

temperature which is the limiting factor for high temperature operations (Munoz-

Anton et al., 2014). The hot receiver heats up the heat transfer fluid which flows

through it and transports the thermal energy to where it is needed. In most cases

superheated steam is created in steam generators, usually for power generation

and sometimes for process heat.

Figure 7: (a) Schematic of a parabolic trough collector (Cabrera et al., 2013). (b) Actual

parabolic trough from Sunray Energy facility in Daggett (Sun & Wind Energy, 2017).

High temperature parabolic troughs have single axis tracking systems, which

follows the sun during the day, allowing the parabolic mirrors to reflect the DNI

onto the receiver tube. The tracking system can either be orientated on a north-

south axis or an east-west axis. The north-south tracking axis is used to track the

sun along the day, while the east-west axis can be used for seasonal tracking. A

north-south orientation allows for more energy to be collected during the summer,

while the east-west orientation has an even spread over summer and winter

(Baharoon et al., 2015). The solar live steam generation integration option may

benefit from the east-west tracking axis system, as most sugar mills only function

during the crushing season and with the winter months coinciding with most of

the crushing season.

(a) (b)


17

The receiver tube onto which the DNI is reflected and through which the heat

transfer fluid flows is usually put inside an evacuated glass tube for high

temperature applications, similar to the evacuated tube collector. This minimises

convection and radiation losses as mentioned before. For parabolic trough

collectors expansion bellows are fitted at the end of the receiver tube to

accommodate for differences in thermal expansion between the metal receiver and

glass tube (Fernandez-Garcia et al., 2010).

Synthetic oils are by far the most common heat transfer fluid used in parabolic

troughs (Gunther et al., 2013). This is mainly due to its low freezing temperature

(12 °C), relatively high specific heat capacity and the fact that it can be obtained

in large quantities making it easy to operate in parabolic troughs. For locations

with sustained low temperatures through the winter months, trace heating might

be necessary to ensure the oil does not become too thick and reach the minimum

temperature of 12 °C. However, for subtropical climates such as where the sugar

mills are located, trace heating should not be necessary, as temperatures rarely

drop so low, and usually not for very long (Solargis, 2017).

Synthetic oil can only operate up to 400 °C, with almost all operations limiting the

temperature to 393 °C as not to take any risks and damage the oil. Other

disadvantages of synthetic oil are that it degrades over time, it is quite expensive,

it is not environmentally friendly and it presents a fire risk since it is flammable

(Heller, 2013).

In order to increase the solar field’s outlet temperature, the use of molten salt as

heat transfer fluid has been researched. Molten salts can operate at temperatures of

up to 600 °C and holds a lot of advantages for solar thermal power generation.

The main advantage of higher temperatures is a higher Rankine cycle efficiency

(Giostri, et al., 2012), which would allow for more power to be generated.

However, there are few process heat applications that can benefit from these high

temperatures.

Molten salt is widely considered to be the best sensible heat storage medium, with

80 % of CSP plants under construction having molten salt storage systems (Liu, et

al., 2016). If it can be used in the solar field as well as the storage system, direct

storage can be implemented. Previously a heat exchanger was needed between the

synthetic oil of the solar field and the molten salt of the storage tanks. Another

advantage of the higher temperature is that it will result in smaller storage systems

(Gunther et al., 2013). Other advantages of using molten salt are that it is less

expensive than synthetic oils, it is more environmentally friendly and not

flammable.

The biggest worry when working with molten salts is its high freezing

temperature, which can be between 120-220 °C, depending on what salt is used.

This is especially problematic for line focussing systems like parabolic troughs

where there are long pipe lines, increasing the risk of solidification during times


18

of low solar irradiance and low temperatures. There are solutions like trace

heating or the circulation of stored hot salts, but this would result in higher costs,

higher auxiliary electricity consumption by the plant and higher heat losses

(Heller, 2013).

Direct steam generation in parabolic troughs have also been a topic of study for

some while now. Direct steam generation differs from the previously mentioned

heat transfer fluids, because no heat exchanger or steam generator is necessary to

transfer the thermal energy to the power block. Direct steam generation allows for

higher temperatures and no expensive heat transfer fluid is needed (Giostri, et al.,

2012).

Operating a direct steam generation parabolic trough field is a lot more complex

than operating the heat transfer fluids mentioned above, due to the two-phase flow

that exists in parts of the solar field. This requires a sophisticated control system,

furthermore, the mass flow of the water/steam must be high to avoid stratified

flow. Another drawback is that for power generation the solar field has to operate

at the same pressure as the turbine inlet, placing a lot of stress on the solar loop

(Giglio et al., 2017).

Linear Fresnel Collector:

Linear Fresnel collectors are classified as a line focussing technology, like the

parabolic troughs, but uses multiple long, flat mirrors to reflect the sun onto an

absorber above them as in Figure 8 and b. Another difference is that the absorber

is stationary, in a parabolic trough setup the absorber moves with the trough as it

tracks the sun; while for the linear Fresnel collector the absorber is stationary and

only the mirrors move as they track the sun (Liu, et al., 2016). Most high

temperature collectors have a secondary reflector around the absorber which

ensures that all of the radiation reflected upwards from the mirrors eventually hit

the absorber. The absorber is then also encased with a glass cover in front and

insolation at the back.

Figure 8: (a) Reflection of sunrays onto absorber in linear Fresnel collector (Electromagnetic

Foundations of Solar Radiation Collection, 2017). (b) Linear Fresnel collector at Kimberlina,

U.S.A. (CSP World Organisation, 2015)

(a)

(b)


19

Linear Fresnel systems are quite new in terms of commercial deployment, the

National Renewable Energy Laboratory (2017) lists only 15 Linear Fresnel power

projects of which 7 are still under construction. The database for applications of

solar heat integration in industrial processes (AEE Intec, 2017) only lists 2 linear

Fresnel systems used for process heat. Blackdot Energy (2017) also lists a linear

Fresnel system which is used in an absorption cooling system in Johannesburg,

South Africa.

Linear Fresnel systems are seen by many as a technology which can significantly

reduce the price of CSP. The collectors cost less to produce because the mirrors

are cheaper than that of parabolic troughs, the structure is lighter and it is less

affected by high wind speeds (Abbas et al., 2016). Direct steam generation can

further reduce prices by eliminating the need for expensive heat transfer fluids.

However, the collector efficiency of linear Fresnel collectors is lower than for

parabolic trough collectors. This is due to lower optical efficiencies of the mirrors

and because of blocking and shading of mirrors. Parabolic troughs also have

higher thermal efficiencies due to the absorber being enclosed in a vacuum.

Abbas et al. (2016) report that linear Fresnel collectors are 37 % less efficient than

parabolic troughs, but that they can generate 21 % more energy per square meter

of land used due to its more efficient use of space. Although linear Fresnel

systems are not the most efficient and technical CSP option, it does offer various

economic benefits and can become the best commercial choice in the right

conditions (Peterseim et al., 2014).

Central Receiver Systems:

The first commercial central receiver power plants were only built in 2007, near

Seville in Spain. Making it a relatively new commercial CSP option, although the

first test facilities have already been built in 1978, near Albuquerque, in the

U.S.A. (Ho, 2017). Central receiver systems or power towers, as they are often

called, consists of an elevated receiver onto which an array of mirrors (called

heliostats) reflects and focusses the sun on, as can be seen in Figure 9. The

heliostats follow the sun on a two axis tacking system in order to focus it on the

central receiver, making it a point focussing system (Stine and Geyer, 2001).

Central receiver systems have the potential to reach much higher temperatures

(above 1000 °C) than parabolic trough and linear Fresnel collectors (Liu, et al.,

2016). This is due to the much higher concentration ratio that can be reached

through the heliostat field. As mentioned before, there are various advantages

related to higher temperatures in solar power plants. This is why 60 % of the

planned CSP power plants at the end of 2016 were central receiver systems (Ho,

2017). At the moment central receivers are mainly being used for power

generation, since this application can utilise the high maximum temperatures.


20

Figure 9: Central receiver system at Crescent Dunes (Solar Reserve, 2017).

There are various heat transfer fluids and mediums which can be used in central

receivers like air, gas, ceramic particles and liquid sodium which can handle the

high temperatures; but most installed systems still use water/steam and molten

salts as heat transfer fluids, operating at temperatures below 600 °C (Ho, 2017).

The design of receiver itself depends what heat transfer fluid is being used, to

ensure efficient heat transfer and minimal radiation heat losses. For the very high

temperature systems it is very important that the receiver does not experience

extremely high radiative losses due to the high temperatures, as this would defeat

the purpose of trying to create a more efficient plant.

The high temperatures suit power generation perfectly, but process heat

applications cannot really benefit from it. This might be the reason why there are

no recorded central receiver systems which are used for solar process heat,

according to the database for applications of solar heat integration in industrial

processes (AEE Intec, 2017).

One drawback of using a central receiver system in a subtropical climate is that

the high humidity associated with these areas would negatively impact the solar

field efficiency. This is because the radiation reflected from the heliostats are

absorbed and scattered by the moisture in the air as it travels to the receiver

(Cardemil et al., 2013). The solar attenuation due to humidity and aerosol can

reduce the annual plant yield up to several percentile points, depending on the size

of the solar field, location and the plant’s operation strategy (Hanrieder et al.,

2017).

3.1.2 Solar resource review

Solar thermal technologies utilises the radiation within the ultraviolet, visible and

infrared spectrums from the sun to supply thermal energy (Kalogirou, 2009). The

sun emits radiation at an intensity of 6.33 × 107 W/m2 from its surface (Stine

and Geyer, 2001), the intensity of the radiation decreases with distance from the


21

sun. The radiation falling on Earth’s outer atmosphere varies between 1330 W/m2

and 1400 W/m2, depending how far the Earth is from the sun during its elliptic

orbit (Stine and Geyer, 2001).

The solar radiation reaching the earth’s surface is usually lower than the radiation

which reached the outside of the atmosphere. This is due to gas, dust and water

particles in the sky which absorbs some of the radiation and also reflects and

scatters it (Duffie and Beckman, 2006). Some of the reflected and scattered

radiation leaves the earth’s atmosphere, but most of the scattered irradiance

eventually reaches the surface due to the entire sky vault, this is called diffuse

radiation (Stine and Geyer, 2001). The solar radiation that directly reaches the

earth’s surface is called direct or beam radiation.

Concentrating collectors only use direct irradiance, as diffuse irradiation cannot be

concentrated due to its nature of coming from various angles to the collector.

Non-concentrating collectors can use both direct and diffuse irradiance. In order

to measure how much irradiance is available to concentrating tracking collectors,

the direct normal irradiance (DNI) is measured. As can be seen in Figure 10.a, this

is the direct radiation that will reach the collector when the collector is directly

facing the sun.

Figure 10: (a) Direct normal irradiance. (b) Global tilted irradiance. (Meyer, 2016)

To measure how much irradiance is available for the stationary, non-tracking

collectors, the global tilt irradiation (GTI) is measured. As can be seen in Figure

10.b, this is the total radiation that reaches the tilted surface. The surface is tilted,

as most collectors are installed at an angle to the horizontal, equal to the latitude

of the location. This will enable the collectors to get a relatively equal distribution

of power throughout the year. Using this tilt, the sun’s angle at noontime will only

vary by a maximum 23.5 degrees, above or below the normal, minimizing the

cosine losses (Stine & Geyer, 2001). The most common way, however, to assess

the solar resource at a location is to look at the global horizontal irradiance (GHI),

it is very similar to the GTI, but the plane considered is horizontal.

(a) (b)


22

Beukes (2015) and Hess (2016) assessed the solar resource at various sugar

milling locations. They found that Durban would make a good reference location,

as it has a similar solar resource compared to the other sugar milling locations.

The GTI only varies 5 % between milling locations, while the DNI varies 10 %.

Figure 11 shows a solar resource map of KwaZulu Natal, showing the annual DNI

on the left and the annual GTI on the right. From these maps we can come to

expect that the annual DNI at Durban should be about 1250 kWh/m2 to

1400 kWh/m2, while the GTI can be between 1700 kWh/m2 and 1850 kWh/m2.

Figure 11: KwaZulu Natal Solar Resource (GeoSUN, 2012).

One of this study’s objectives is to create detailed simulations of the solar thermal

systems. In order to do this, accurate data will be needed, therefore solar data was

purchased from Solar Resource Data Solargis©. The purchased data is hourly

typical meteorological year (TMY) solar data. This set of data consists of 12

months selected from individual years to form a year of data which best represent

the average solar radiation and weather conditions (Solargis, 2017).

The TMY data purchased for this study stretches over a time span of 22 years.

The data gives the GHI, GTI and diffuse irradiance of Durban, by using this it was

also possible to calculate the DNI with Equation 3.1 as recommended by

Meyer (2016). Here h represents the sun’s elevation angle.

𝐷𝑁𝐼 = 𝐺𝐻𝐼−𝐷𝑖𝑓𝑓𝑢𝑠𝑒

sin (ℎ) (3.1)

It was calculated that Durban’s annual DNI is 1350 kWh/m2, falling in the

expected range, mentioned previously. The GHI is 1609 kWh/m2 and the GTI is

1821 kWh/m2, at the upper limit of what was expected. This data shows that

Durban is better suited for collectors utilizing GTI, like the flat plate and


23

evacuated tube collectors which can be utilized for the bagasse drying sysem, as

there will be a higher solar resource available to them.

Looking at Durban’s TMY DNI data, it can reach values above 1000 W/m2 quite,

regularly. However, Durban’s tropical climate causes lots of clouds to form

during the day and from September to November there tends to be lots of rain,

which blocks off DNI. Figure 12 shows a comparison of the DNI of Durban and

Upington for 21 days. From this figure, one can see that the total amount of DNI

for Durban is significantly lower than for Upington, due to transient conditions.

This is why the region around Upington is preferred for the development of CSP

plants. It is thus expected that the solar live steam generation would not perform

as well as a commercial CSP plant, but would still be able to make a significant

impact due to the high maximum DNI which can be reached quite regularly.

Figure 12: DNI comparison of Upington and Durban (Meyer, 2016)

As part of the purchasing agreement, Solargis© compiled a report of the solar

resource in Durban. Figure 13.a shows how much the GHI varied for the 22 years

the data was measured and highlights the importance of using averaged or TMY

data rather than the data of a chosen year. Figure 13.b shows the monthly averages

of GHI over these years.

Figure 13: (a) Annual GHI for Durban from 1994 - 2016. (b) Monthly average GHI for

Durban (Solargis, 2017)

(a) (b)

Hours

DN

I [W

/m2 ]


24

3.2 Bagasse Drying

Literature on bagasse drying is analysed to highlight the importance of the

process. Potential problems and drawbacks found in the literature are also

discussed here. The drying process itself will be explained and the different types

of bagasse dryers will be analysed.

3.2.1 Advantages of bagasse drying

When bagasse leaves the dewatering mills it is still rather wet, about half the

weight of bagasse consists of water. The average moisture content of the bagasse

fed into the boilers across the South African industry was 50.78 % (weight

percentage) for the 2015/2016 milling season (Smith et al., 2016) and in the

BRTEM Matlab model bagasse moisture was set to 51 % (Starzak and Davis,

2016). As mentioned before, one of the main advantages of bagasse drying is that

it can significantly increase the calorific value of the bagasse. By increasing the

calorific value, less bagasse has to be used to produce the same amount of heat.

The higher heating value, also known as the gross calorific value (GCV),

represents the total energy that can potentially be released by bagasse when

combusted (Wienese, 2001). If there is no moisture, ash or Brix content in the

bagasse, this value is 19 605 kJ/kg (Don et al., 1977). The moisture, ash and Brix

contents have a negative impact on the calorific value. Don et al. (1977)

developed a formula for determining the GCV, shown as Equation 3.2. M

represents moisture content, B the Brix content and A the ash, all expressed as

weight percentages of the bagasse.

GCV = 19605 – 196.05 M - 31,14 B – 196.05 A (3 .2)

The lower heating value, also known as the net calorific value (NCV) represents

the amount of energy that the combustion of bagasse can release minus the latent

heat of the water formed during the combustion process. The NCV is a good

indication of the heat that is theoretically available from the combustion of the

fuel. It also gives a better indication of what is realistically possible (Hugot,

1972). Equation 3.3 from Wienese (2001) shows how to calculate the NCV:

NCV = 18260 – 207.01 M – 31.14 B – 182.60 A (3.3)

From both Equations 3.2 and 3.3 it is clear that reducing the moisture content in

the bagasse will lead to an increase in calorific value. A reduction of 5 % in the

moisture content can lead to a 12 % increase in GCV (Vijayaraj et al., 2007).

Sosa-Arnoa et al. (2006) mentions a study where bagasse was dried from 50%

moisture content to 38% and this resulted in an a 16% increase in steam

production. Figure 14.a shows how bagasse drying improves the fuel usage to

steam production ratio.


25

Figure 14: (a) Bagasse to steam ratio as a function of the bagasse moisture content. (b) Boiler

fan power as a function of the bagasse moisture content (Magasiner, 1987).

Another advantage of bagasse drying is that the boiler efficiency increases; this is

because by reducing the water vapour released by the bagasse during combustion

the heat losses through the flue gas can be reduced. Furthermore, dried bagasse

needs less excess air for effective combustion, this, along with the reduced

amount of vapour will lessen the load of the induced draft fans (Sosa-Arnoa et al.,

2006; Bruce & Sinclair, 1996). Figure 14.b shows the effect of bagasse moisture

on absorbed fan power.

Older boilers with old types of grates like pin hole or steeply sloping fixed grates,

sometimes struggle with handling wet fuel and could benefit greatly from a

bagasse drying operation (Wienese, 2001). This is because the moisture content in

bagasse has a significant effect on the combustion process itself. Dryer fuel

improves combustion since it burns faster and hotter. This in turn enables the

boiler to be more responsive to changes in load (Bruce & Sinclair, 1996).

3.2.2 Potential problems and disadvantages of bagasse drying

Laubscher (2017) warns that a major reduction in combustion air, because of

dryer bagasse, may lead to a smaller convection heat flux in the boiler, leading to

lower steam temperatures and boiler efficiency. However, for the bagasse drying

up to 35 % moisture content this will not be an issue as Figure 15 shows. Here we

can see that the boiler’s efficiency still increases as the bagasse moisture is

reduced, however, the increase is slightly less for each 5 % moisture content

interval.

(a) (b)


26

Figure 15: Simulation of different bagasse moisture contents for a boiler, similar to the one

assumed for the BRTEM model (Laubscher, 2017)

The hotter and more vigorous combustion of dry biomass compared to wetter

biomass increases the probability for ash fouling and slagging. Fouling occurs

when volatile ash components (usually alkali metal oxides) condenses in the

boiler and attaches itself to the heat transfer surfaces, where it is then later sintered

due to the extreme heats (Magasiner et al., 2001). Fouling is more likely to occur

with higher concentrations alkali metals in the fuel and there are various ways to

try and determine if it will occur. A simple measure to see if there is a risk for

fouling, is to look at the mass ratio of alkali metal oxides to silica, as shown in

Equation 3.4.

(Na2O + K2O) : SiO2 (3.4)

Bruce & Sinclair (1996) gives and index that can be used as a rough guide to predict

fouling using Equation 3.4. According to this index a ratio above 2 indicates a high

fouling potential. Miles, et al., (1995) developed a more complex fouling indicator

formula, as shown in Equation 3.5.

1×106

𝐺𝐶𝑉𝑘𝐽

𝑘𝑔(𝑑𝑟𝑦)

× 𝐴𝑠ℎ% × 𝐴𝑙𝑘𝑎𝑙𝑖% 𝑖𝑛𝐴𝑠ℎ = 𝑘𝑔𝐴𝑙𝑘𝑎𝑙𝑖

𝐺𝐽 (3.5)

For a value higher than 0.17 kg/GJ the risk of fouling increases and for a value

higher than 0.34 kg/GJ fouling is almost certain to occur. Magasiner et al. (2001)

suggests that this should be used together with Equation 3.4 above, since Miles’s

indicator may sometimes overstate the potential problem. This happens when the

ratio is moderate or high but the total amount of the offending materials is too

small to have a significant effect. For Equation 3.4 bagasse is calculated to have a

ratio of 0.06 and for Equation 3.5 the answer is 0.07 kg/GJ, both of the values are

far below the predicted fouling points.

86

86,5

87

87,5

88

88,5

89

89,5

90

90,5

91

40 60 80 100 120

NC

V e

ffic

ien

cy (

%)

Boiler load (%MCR)

35%Moisture

40%Moisture

45%Moisture

50%Moisture

55%Moisture


27

It is, however, important to keep in mind that these indexes are only indicators of

potential problems. Fouling cannot be anticipated by looking at fuel properties

alone. Although the tendency for this generally increases with higher alkali

content in the fuel, the form of the alkali and other inorganic constituents along

with boiler operating conditions and boiler design play an important role as well

(Miles, et al., 1995).

Slagging can occur when the high temperatures in the furnace reaches a level

where the ash begins to exhibit stickiness. This enables the ash to attach itself to

colder surfaces and prohibit effective heat transfer in the boiler (Bruce & Sinclair,

1996). The potential of slagging occurring with a certain fuel can be determined

by looking at the phase diagrams to see if the boiler’s operation temperature

would melt the ash or not (Magasiner et al., 2001).

The ash fusion temperature of bagasse is 1310 °C – 1380 °C according to

Magasiner et al. (2001). This is the temperature where ash becomes soft and starts

to melt. This temperature, however, can be lower when bagasse is used in multi-

fuel firing. It is possible that ashes from the different fuels chemically interact to

form a eutectic compound with a lower ash fusion temperature than that of the

ashes of the initial fuels (Rayaprolu , 2009).

However, bagasse has been used successfully as a boiler fuel without any major

concerns, even when used with coal. This could be attributed to the fact that both

potassium and chlorine are substantially leached from the sugarcane in the process

of sugar extraction (Miles, et al., 1995).

When dryer bagasse is used it is important to take the boiler design into

consideration (Magasiner, 1987). If a boiler was designed for high moisture

content bagasse, then the bagasse should not be dried too much, since this might

create complications in boiler operations. For example boilers with a refractory

band in the ignition zone can experience heavy fouling if bagasse moisture levels

drop below 45 %. It would be best to remove the refractory band if the bagasse

fed into the boiler is predominantly below 45 % in moisture content. It would be

better to use refractory backed open pitched tubing in the ignition zone

(Magasiner, 1987). The SMRI also feels that substantial changes may have to be

made to current South African sugar milling boilers if it uses bagasse with a

moisture content below 40 % (Foxon, 2017).

Another aspect that has to be kept in mind when drying bagasse is what will

happen if bagasse with different moisture contents are fed into the boiler. In order

to do this it is important to mix the bagasse so that there is an even spread of the

wetter and dryer bagasse and a constant fuel density (Naude et al., 1993). This

will prevent the formation of puffs in the boiler. Furthermore, it is important to

control the fuel-air ratio to ensure efficient combustion (Naude et al., 1993).


28

Rein (2007) also notes that a bagasse drying system adds additional cost and

complexity to a sugar mill. The extra technology needed to support a drying

system, apart from the dryer itself, like fans and a conveyor system can add

significant additional cost.

3.2.3 Drying process

During the drying process the liquid on a solid’s surface will be heated to an

equilibrium temperature by the hot air stream. At this temperature the rate of

drying is constant due to the fact that the evaporation of moisture absorbs latent

heat. This can be seen as the horizontal line between points B and C on Figure 16.

During this constant drying rate period, the moisture movement inside the solid is

rapid enough to maintain a saturated condition at the solid’s surface. Eventually

the moisture content has been so reduced that dry spots start to appear on the

surface, this reduces the drying rate quite significantly as can be seen by the line

between C and D in Figure 16. With further drying the drying rate starts to depend

on the rate at which moisture can move through the solid, as a result of

concentration gradients (Tawfik et al., 2003). This period is represented by the

line between points D and E in Figure 16.

Figure 16: Drying rate under constant drying conditions (Tawfik et al., 2003)

However, some studies (Vijayaraj et al., 2007; Bakshi& Singh, 1980; Freire et al.,

2001) have shown that bagasse does not quite follow this drying curve. Bagasse

has no constant drying rate period, despite having relatively high moisture

content. Freire et al. (2001) observed that the bagasse goes through a short heating

period and from there on dries only according to the falling rate period, points C

to E in Figure 16. Vijayaraj et al. (2007) mentions that bagasse’s drying rate is

mainly dependant on the moisture diffusion inside the bagasse.


29

When drying bagasse the moisture is released between 120 °C and 210 °C. When

the bagasse is heated to more than 210 °C it starts to emit CO2 and at 228 °C it

starts to combust (Naude et al., 1993). This is clearly evident in Figure 17, which

shows how the CO2 emissions of bagasse increase when it reaches a temperature

above 200 °C and also how it starts to lose weight a short while after that. Sosa-

Arnoa and Nebra (2009) recommend that the hot air temperature should not

exceed 200 °C by much. They used 215 °C for their experimental setup, which

uses a pneumatic dryer. Fires do not occur often in drying systems, but most of

those that have occurred started in the collector due to particle build-up, which

creates favourable circumstances for fires (Cook, 1991).

Figure 17: The heating of bagasse under an oxygen atmosphere from Sosa-Arnoa and Nebra,

2009)

High temperatures are beneficial for the drying system since it increases heat

transfer and minimizes equipment size. However, it does create a fire hazard. The

fire risk can be minimized by understanding the different drying phases (Amos,

1998). There are two points where the fire risk is quite high. One point is where

dry spots start to form (point C in Figure 16). During this period there is no water

vapour on the particle’s surface which causes the particles’ temperature to rise to

dangerous levels. So bagasse might be under constant risk if we assume that it has

no constant drying rate interval. However, the moisture which is still within the

particle helps to control this risk since it moves to the surface, which in effect

cools the particle again (Amos, 1998).

The second high fire risk arises when the particle is completely dry. The particle’s

temperature would rise until its combustion temperature, if the drying air is hot

enough. This is luckily a very small risk since bagasse is never dried to this point;

as mentioned before, bagasse has a technical lower limit for drying in the order of

30 % (Rein, 2007).


30

3.2.4 Dryer types

Sosa-Arnoa et al. (2006) did a review of bagasse drying and of the 32 drying

systems that they listed, 20 was pneumatic dryers. The other 12 was mainly rotary

dryers. This study will only look at these two dryer technologies, since they are

the technologies that are most used in the specific industry. Pneumatic dryers are

most often used for bagasse drying because they are relatively cheap and have

small space requirements (De Oliveira, et al., 2011). Rotary dryers are more

robust than pneumatic dryers and can handle the largest capacity of any type of

dryer; it can also be used in cases where there are large particles or a variety of

particle sizes (Amos, 1998; Bruce and Sinclair, 1996).

Rotary Dryers:

In a rotary dryer, the hot air comes in contact with the bagasse inside the rotating

drum. Most of the drying takes place as flights inside the drum lifts the bagasse

and lets it cascade through the hot air stream (Cook, 1991). This allows more of

the bagasse particles to be in direct contact with the air and promotes mass and

heat transfer.

The flow of hot air and bagasse can be co-current or counter current. For the latter

case the driest solids are exposed to the highest temperatures with the lowest

humidity. This can be seen as a possible fire risk and would not be advised for

drying systems working at high temperatures (Amos, 1998). Counter current

drying can be problematic when working with bagasse, because bagasse particles

are light and have high drag coefficients and therefore a lot of the particles can be

caught in the airstream, blowing it back to the inlet (Foxon et al, 2017).

Co-current rotary dryer systems are designed so that the wettest particles come in

contact with the hottest air, minimizing the fire risk. Some systems are designed

so that the exhaust stream passes through an air and fine particle separator like a

cyclone, baghouse filter, scrubber or electrostatic precipitator to retrieve any

particles still entrained in the air stream, as shown in Figure 18.

The residence time for very small particles can be as short as 30 seconds, but for

most of the bigger particles in is in the order of 10 – 30 minutes. This long

residence time allows the particles to dry out quite uniformly. Non-uniformity in

particle moisture content can be a problem with certain dryers, especially if there

is a large range of particles sizes present (Bruce & Sinclair, 1996). The bagasse

should typically occupy 9 – 15 % of the drum shell to ensure effective drying

(Cook, 1991) and the dryer is most efficient with length – diameter ratio between

4-10 (Mujumdar et al. 2006). One of the drawbacks of a rotary dryer system is

that the moisture content of the particles are hard to control because the rather

long residence time create a lag in the system (Fredrikson, 1984).


31

Figure 18: Co-current rotary dryer system

Table 2 shows some of the capabilities of a rotary dryer. The bagasse drying

operations being considered for this study fall well within the feed moisture

content and discharge moisture content performance parameters of the rotary

dryer. However, the particle sizes of the bagasse do not quite match that of the

rotary dryer’s design. Bagasse has a wide variety of particle sizes, which can

range from 1 mm to 150 mm. 30 % of the bagasse consist of very small particles,

called pith, which forms part of the lower range of sizes, usually ranging from

1 – 5 mm ( Foxon et al., 2017; Rein, 2007).

Table 2: Typical design and performance values for rotary dryers (Bruce & Sinclair, 1996)

Parameter Unit Value

Evaporation t/h 3 – 23

Capacity t/h 3 – 45

Feed moisture content % 45 – 65

Discharge moisture content % 10 – 45

Pressure drop kPa 2.5 - 3.7

Particle size mm 19 – 125

Thermal requirements GJ/tevap 3 – 4

Pneumatic Dryers:

In pneumatic dryers, also called flash dryers, the hot air carries the solid particle

up a flash tube, drying the material through direct contact, Figure 19 shows a

typical pneumatic dryer setup. These type of dryers have short contact times

between the hot air and wet material, usually between 0.5 – 10 s (Mujumdar et al.,

2006). This is due to the large surface area of the material exposed to the air,


32

where heat and mass transfer can take place. The short contact time minimizes the

fire hazard because the material temperature stays relatively low.

Figure 19: Pneumatic dryer system

In order for the hot air stream to transport the material the air velocity needs to be

larger than the free fall velocity of the largest particle to be dried (Mujumdar et

al., 2006). Strumillo and Kudra (1986) recommends that the air velocity should be

2.5 m/s above the largest particles’ terminal velocity. Therefore, in a pneumatic

drying system the ratio between the air velocity and particle velocity needs to be

high. Table 3 shows some of the capabilities of a conventional pneumatic dryer.

From these values it is possible to see that pneumatic dryers can dry of more

moisture than rotary dryers and that they use less heat, however the typical

discharge values are considerably lower than what is considered for this study.

Pneumatic dryers, furthermore, are designed to dry a smaller range of particles

sizes compared to rotary dryers.

Table 3: Typical design and performance values for a pneumatic dryer (Bruce & Sinclair,

1996)

Parameter Unit Value

Evaporation t/h 4.8 – 17

Capacity t/h 4.4 – 16

Feed moisture content % 45 - 65

Discharge moisture content % 10 - 15

Pressure drop kPa 7.5

Particle size mm 0.5 - 50

Thermal requirements GJ/tevap 2.7 – 2.8


33

Pneumatic dryers are simple to construct and have low capital cost. The

maintenance cost is also quite low due to the small amount of moving parts.

However, the operating cost of pneumatic dryers is high because of the blower

system that needs to maintain a high air flow rate through the system (Amos,

1998).

3.3 Cogeneration and Hybridisation with CSP

By using biomass along with CSP, electricity can be generated using renewable

energy at low CO2 emission levels (Burin, et al., 2016). CSP-biomass hybrid

power plants can be used as reliable base load providers, the only problem being

that there are only a few areas with both adequate solar and biomass resources

(Peterseim et al., 2013). Peterseim et al. (2013) mentions that possible locations

should have an annual DNI higher than 1700 kWh/m2 to be financially feasible.

One of the few suited areas is Les Borges Blanques in Spain, where the first

commercial solar-biomass power plant, Borges Termosolar, has been in operation

since December 2012. The hybrid power plant has a capacity of 22.5 MW and

operates 24 hours a day (National Renewable Energy Laboratory, 2013). The

plant generates 98 000 MWh annually, which is 89 % of the output of a 50 MW

solar-only plant (Biomass Knowledge Portal, 2015).

The solar loop, parabolic troughs with a combined aperture area of 183 120 m2,

heats thermal oil to 393 °C which is used to create steam at 40 bar in the steam

generator. From there on the steam is superheated to 520 °C by the dual biomass-

natural gas boiler and enters the high pressure steam turbine as can be seen in

Figure 20. The biomass and auxiliary natural gas boiler are tasked with heating

the thermal oil during times of low solar irradiance. They hybridisation helps to

save biomass and allows the steam turbine to operate without any interruptions,

resulting in higher efficiencies.

Figure 20: Layout of the Borges Termosolar hybrid power plant (Biomass Knowledge

Portal, 2015).


34

The way a hybrid plant operates can differ from plant to plant, depending on what

they want to achieve and different economic factors. Peterseim et al. (2014)

completed a study which evaluates different hybrid plant designs for Australia. In

his design the biomass boiler provides a constant supply of energy, seen as

Energy from Boiler (EfB) or Energy from Waste (EfW) in Figure 21. The CSP

system provides additional power during daytime, when the electricity demand

and prices are higher in Australia.

Different CSP technology can be implemented with the hybridisation, this will

have an impact on the solar system efficiency as well as on the economic factors

like the IRR. The studies of Peterseim et al. (2014) and Iftekhar Hussain et al.

(2017) found that a linear Fresnel system with direct steam generation would yield

the best economic results in a hybrid plant compared to parabolic trough, solar

tower and other linear Fresnel systems with different heat transfer fluids.

Figure 21: Electricity output of the hybrid concept designed by (Peterseim et al, 2014)

Adding CSP to an existing biomass plant would be cheaper than developing a

stand-alone CSP plant, due to the fact that infrastructure can be shared. Most of

the savings come from the power block, where most biomass plants and sugar

mills already have boilers, turbines, generators and condensers (Bhatt, 2014).

Peterseim et al. (2014) reports that the investment costs of a new CSP and

biomass hybrid plant is 12 % lower compared to a stand-alone CSP power plant,

also mainly due to the fact that infrastructure can be shared. On top of that, a

hybrid plant can generate more power annually than a stand-alone CSP plant of

the same rated capacity. In order to create a CSP plant to generate the same annual

electricity output, the investment costs will have to be 69 % higher than that of the

equivalent hybrid plant (Peterseim et al., 2014). A hybrid plant, is however, more

expensive than a biomass-only plant, but as biomass prices increase, it can soon

be cost competitive with the biomass-only plants (Nixon et al., 2012).


35

As noted in Section 2, sugar mills (without refineries) usually only operate during

the crushing season. The installation of CSP can enable the power cycle in the

mills to operate outside of the crushing season as well, increasing the amount of

electricity exported per annum. Sugar mills are an important source of electricity

in large sugar producing countries. Brazil, for example, has 10.6 GW installed

electrical capacity at sugar mills, which represents 7.1 % of the country’s total

installed capacity. Sugar mills can benefit from CSP integration in the same way

as biomass plants. It will allow for biomass to be saved and extra electricity can

be generated without having to make capital investments into the power block.

Burin et al. (2016) completed an evaluation of integrating a solar power tower into

an existing Brazilian sugarcane cogeneration plant. The study evaluated three

different solar integration points, similar to Hess et al. (2016). The integration

points were: feed water preheating, saturated steam production and live steam

production. The live steam production proved to have the biggest impact, with the

other two integration points proving costly in terms of the impact that they could

make.

In the live steam production integration point, shown in Figure 22, CSP was used

to create live steam during sunny hours, easing the boiler’s load. Outside of the

crushing season the CSP system can run on a solar only mode during sunny hours,

utilising the condensing extraction steam turbines (CEST) installed at the mill to

create electricity. The bagasse saved during the crushing season can be used

during hours of low solar irradiance. According to Burin et al.’s study this

hybridisation mode can increase electricity production by 19.8 %. These results

show that a CSP system can still make a considerable impact despite having

annual DNI readings lower than 1700 kWh/m2 as suggested by Peterseim et al.

(2013), since the annual DNI reading for the site considered by Burin et al. (2016)

is 1502 kWh/m2.

Figure 22: Simplified layout of CSP integration into a Brazilian sugar mill (Burin, et al.,

2016)


36

The bagasse savings fulfils the role thermal storage would normally serve,

allowing hybrid plants to save on investment costs (Servert et al., 2011).

However, storage can increase the percentage of electricity which the CSP system

itself delivers annually. The increased electricity production, reduction in biomass

usage and extra capital costs almost balances each other out, with CSP systems

utilising thermal storage only showing a slight drop in IRR (Peterseim et al.,

2014).

Soria et al. (2015) shows that in order for a solar system to deliver 50 % of the

needed energy, the ratio of installed thermal capacity of CSP to biomass should be

70:30. The biomass needs less installed thermal capacity since it can run 24 hours

a day, whilst the CSP system can only run during sunny hours. The solar multiple

can also be increased to allow for a higher annual CSP capacity factor. The solar

multiple is defined as the ratio between the thermal power which the solar field

could deliver at design conditions and the amount of thermal power needed at

nominal conditions (Montes et al., 2009). So higher solar multiples allows for

more thermal energy, but during times of high irradiance it will be too much,

forcing parts of the solar field to defocus. Defocussing lowers the solar field’s

efficiency, due to the fact that it does not use the energy available.

Burin et al. (2016) found that a solar multiple of 1.4 is ideal for the designed CSP

system in their study, resulting in the lowest levelised cost of energy (LCOE) of

220 US$/MWh as can be seen in Figure 23. The LCOE is the average price per

unit of energy generated by the system (Hess et al., 2016). It is a common way to

assess the feasibility of a project and is frequently used to compare renewable

energy projects to one another.

Figure 23: LCOE and additional electricity produced for different solar multiples (Burin, et

al., 2016)

The LCOE of 220 US$/MWh is in the same range as older commercial CSP

plants, whose LCOEs ranges from 200-400 US$/MWh(Burin, et al., 2016). This is

quite a good result considering the low annual DNI mentioned previously.


37

However, CSP plant have reached record low prices in 2017, with SolarReserve

reaching a price of 61 US$/MWh with its Aurora project in Australia (Kraemer,

2017). Servert et al. (2011) calculated the LCOE of a hybrid plant to be

153 €/MWh or roughly 183 US$/MWh.

The differences in LCOE can be due to a number of factors since the LCOE is

subject to change depending on technology and location. In Servert et al.’s (2011)

study the annual DNI was 2000 kWh/m2, but the factor that probably had the

biggest impact is that Burin et al.’s (2015) study is based on sugar mill

technology, while Servert et al. (2011) and the Aurora project focuses on power

generation technology. Sugar mills focus on co-generation, so a large fraction of

the thermal energy and electricity is used for the milling process, which in most

cases is the first priority, the extra thermal energy can then be used for power

generation and electricity export.

Despite the Borges Termosolar plant running for 5 years, the hybridisation of

solar and biomass is still seen as relatively new and unproven technology (Iftekhar

Hussain et al., 2017). The uncertainty surrounding the technology due to its novel

status means that there are various economic and technological hurdles it still has

to overcome before being widely accepted.

Technical hurdles that need to be overcome are the balancing out the different

fluctuations in each of the solar field and biomass boilers, as well as avoiding

situations where thermal energy needs to be dumped (Iftekhar Hussain et al.,

2017). Some of the economic hurdles are similar to what stand-alone CSP plants

face, like obtaining the necessary funding for the high initial capital costs. Milling

companies are focussed on maximising their profits and prefer to invest in

projects with short payback periods, something solar thermal integration cannot

always offer due to the high capital investment (Backen et al.,2017). An issue

which can arise in South African, is finalising a power purchase agreement with

Eskom, the national power provider, an issue which has proven to be a massive

obstacle for many independent power producers (SAIPPA, 2017).

However, hybrid plants can become a lucrative option in the future due to the

predicted increases of fossil fuels, the expected decrease in CSP installation costs

(IRENA, 2016) and countries’ willingness to decrease carbon emissions. Biomass

prices are also expected to increase (Nixon et al. 2012) and can be used to

generate extra income through the production of bio-ethanol or selling it for other

applications.


38

4 Solar Bagasse Drying

This section will explain how the proposed bagasse drying system will work with

solar heated air. The section will go on to explain how the drying system’s

characteristics and thermal requirements were determined using Simprosys; how

the solar system was modelled using a Matlab code and Aircow; and finally how

the solar field was simulated using Solgain, in order to see how the system would

perform during the crushing season.

4.1 Integration Point

The primary focus of introducing bagasse drying at a sugar mill will be to reduce

bagasse usage. However, dryer bagasse also presents various advantages in the

boiler, as discussed in the literature review. As mentioned in Section3.2, most of

the studies done on bagasse drying uses pneumatic dryers (Sosa-Arnoa et al.,

2006). The flue gas temperatures in South American sugar mills (where most of

these studies were done) tend to be considerably higher compared to South

African sugar mills, this is due to the fact that the whole mill operates at higher

steam temperatures and pressures. Therefore, the high temperatures needed to

quickly and effectively dry bagasse in a flash dryer might not be achievable in

South African sugar mills.

Another issue with using flue gas to dry bagasse in South Africa is the risk of flue

gas condensation, which can cause corrosion due to the acid that can form. In

South Africa flue gas temperatures are typically 160 °C and its dew point

temperature ranges from 90 °C - 152°C, depending on the amount of sulphur in

the gas (Laubscher, 2017; Kotze, 2016). The higher the sulphur content, the

higher the dew point and the larger the risk of condensation, and in South Africa it

is possible to get high amounts of sulphur in the flue gas due to the use of coal.

The risk of using flue gas opens the door for low – medium temperature solar

thermal integration and for the drying system a co-current rotary dryer was chosen

to dry the bagasse. This technology was chosen due its robustness, ability to

handle various sizes of particles and ability to handle large amounts of bagasse at

a time.

A co-current system was chosen over a counter-current system, because of the

risk of small particles of bagasse being transported into the wrong direction with

the counter-current air stream. With a co-current system a cyclone can be placed

at the end of the dryer to collect the smaller bagasse particles which were carried

away with the air stream. The fan is placed at the beginning of the system and not

after the dryer systems, as seen in Figure 24, so that it does not have to endure the

relatively high heat. This will also make place for the cyclone and any other dust

and particle collectors that may be put in place if necessary. Furthermore, the fan

would use slightly less power if installed at the front, as all the evaporated

moisture does not have to pass through it


39

Steamboilers

Live steam

390 °C | 31 bar96.76 t/h

Flue gas

Feed water

113 °C | 31 bar | 96.96 t/h

Bagasse (40 % moisture)

Bagasse (51 % moisture)

Evacuated Tube

Collector

Dry air26 °C | 1 bar

64.5 °C | 1 bar43.63 t/h

Wet air

Solar thermal system Integration Conventional system

(160 - 280 °C)

Dry air120 °C | 1 bar

Rotary Dryer

Exhaust Steam127 °C| 2.5 bar

Condensate

Pre-heated air50°C | 1 bar

Figure 24: Basic schematic of solar bagasse drying integration point.

The system starts by sucking in ambient air through the fan mentioned above. The

air is pre-heated by the air exiting the rotary dryer. The air is then pre-heated even

further by a field of evacuated tube air collectors. Evacuated tube air collectors are

used for this integration point, because they are uncomplicated and its temperature

range falls perfectly in the safe temperature for bagasse drying. Exhaust steam

also falls in this temperature range and is already used as heating medium in the

sugar drying operations installed in various South African sugar mills. The

collector field can heat the air up to any temperature between 50 °C and 120 °C

and the exhaust steam air heater will ensure that the final air temperature is

constantly 120 °C.

The exhaust steam in typical South African sugar mills can range from 117 to

127 °C (Foxon, 2017). The exhaust steam at each factory also does not have to

stay at a certain temperature, as most mills vary between the above mentioned

temperatures between cleaning cycles, using hotter exhaust steam when scaling

reduces the heat transfer capacity in some of the processes. Therefore, although

the BRTEM model lists the exhaust steam temperature as 121 °C, it can easily be

adjusted to 127 °C, which should ensure a high enough temperature difference

between the exhaust steam and the air that needs to be heated.

The exhaust steam will allow for a constant drying rate for the bagasse through

day and night. The solar system will reduce the thermal load of the exhaust steam

air heater, saving exhaust steam and, therefore, bagasse. During night time the

solar loop can be bypassed and the pre-heated air can go directly to the exhaust

steam heater as seen in Figure 24.

It would be of great benefit to the mill if the bagasse was dried to the same

moisture content all the time, since this would greatly simplify boiler operation

and control. If the exhaust steam is not used in the air heating system and the air is

just heated by the solar system, the drying air temperature would change

constantly as the solar irradiance varies throughout the day. This will lead to

bagasse with varying moistures and the plant operators will have their hands full

ensuring the boiler is run correctly. One option to eliminate the fluctuations of the


40

thermal energy produced by the solar field is to implement a thermal storage

system. The solar field’s solar multiple can be increased, with the extra energy

going to the storage system so that the drying system can use it again during

transient conditions.

However, the thermal storage can only supply heat for transient conditions and for

a few extra hours after sunset. It is uncommon for storage systems to be able to

supply the full load thermal energy during night time as well, because this would

require a storage system of considerable size. Therefore, there will always be a

need for the exhaust steam air heater. A storage system will be able to assist the

collector field to offset exhaust steam, but at an extra cost. This study will

therefore not look at thermal storage systems as the energy consumption of the

processes surpasses the capability of the solar thermal systems significantly.

Correct boiler operation is a major source of concern for the sugar milling

industry. This is the reason why the bagasse is only dried to 40 % moisture

content. If the bagasse is dried to below this point the bagasse can respond

differently to suspension burning and most of the South African sugar mills’

boilers are not designed to handle bagasse with very low moisture content.

Another problem arising from bagasse that is too dry is that it can be very hard to

handle, since it tends to fly around, becoming a hazard for people who work with

it (Foxon, 2017). It was due to all the above reasons that it was agreed with the

SMRI to design a bagasse dryer that would dry the bagasse to only 40 % moisture

content.

4.2 Drying Model

4.2.1 Simprosys

Bagasse drying and drying systems in general are very complex. Since it is not the

goal of this study to create a detailed drying simulation, therefore, a simple drying

program was used to model the bagasse drying process. The software package

used is called Simprosys(Simprotek, 2006); it is a Windows based program which

can solve heat and mass balances of drying processes and is based on the work of

Masters (1985), Mujumdar (1995) and Perry (1997). The program can also do

scoping design calculations, which is based on the work of Kemp and David

(2002). Simprosys was designed to help engineers quickly obtain the drying

process parameters and dryer size by calculating the necessary air flow to the

dryer and the heating duty of the air heater.

The program calculates all the necessary psychrometric properties of the air-

vapour mixtures in the drying system for the user. The program then goes on to

use Equation 4.1 and 4.2 from Gong and Mujumdar (2008) as heat and mass

balances respectively. Here �̇�𝑎 is the air mass flow and �̇�𝑏 is the bagasse mass

flow. In Equation 4.1, I represents the enthalpy of the air and bagasse streams,


41

while Qc, Ql, Qt and Qmrepresent the indirectly supplied heat to the dryer, the heat

losses, the heat carried in by the transport device and the mechanical energy input.

In Equation 4.2, Y and X represent the moisture content in the air and bagasse,

respectively, and Mevis the amount of moisture evaporated.

�̇�𝑎(𝐼𝑎,1 − 𝐼𝑎,2) + 𝑄𝑐 = �̇�𝑏(𝐼𝑏,2 − 𝐼𝑏,1) + 𝑄𝑙 + 𝑄𝑡 + 𝑄𝑚 (4.1)

�̇�𝑎(𝑌2 − 𝑌1) = 𝑀𝑒𝑣 = �̇�𝑏(𝑋1 − 𝑋2) (4.2)

The program allows the user to define a generic drying material by entering the

specific heat of the bone dry material. The specific heat is used in Equations 4.3

and 4.4 from Mujumdar (1995). Where cam and cbd represents the specific heat

capacities of the moisture in the air and dry air respectively. The subscripts bm

and bd refer to the specific heat capacities of the moisture in the bagasse and the

bone dry material. The terms ∆hA and ∆hB are latent heat of vaporization and heat

of sorption.

𝐼𝑎 = (𝑐𝑎𝑚 𝑌 + 𝑐𝑎𝑑)𝑇 + ∆ℎ𝑎𝑌 (4.3)

𝐼𝑎 = (𝑐𝑏𝑚 𝑌 + 𝑐𝑏𝑑)𝑇 − ∆ℎ𝑏 𝑋 (4.4)

The above equations form part of Simprosys’s path to determine the temperature

rise of the dried material for a given air flow, or a certain exit temperature can be

assumed for the dried material and from there an air flow can be calculated. When

these equations are balanced, Simprosys can go on to give an estimate of the

dryer’s size.

The program uses Equation 4.5 to determine the cross sectional area (A) of the

dryer. The equation does not differentiate between different dryers; the only way

to do this is to use different gas velocities (vg) as this parameter is different for

each dryer technology. Kemp and David (2002) suggest that 20 m/s be used for

flash dryers, 0.5 m/s for fluidised bed dryers and 3 m/s for co-current rotary

dryers. The length of the dryer depends on the length/diameter ratio, which can be

specified by the user.

𝐴 = 𝑀𝐴

𝜌𝐴 𝑣𝑔 (4.5)

The program does not take the bagasse’s specific drying kinetics or the particle

residence time into account, meaning that it can be over optimistic in some of its

drying results. However, the cross sectional area of the needed dryer can be

estimated surprisingly accurately (Kemp and David, 2002). According to the

program developers, Simprosys is the first step towards a comprehensive drying

suite. It is affordable and user friendly, allowing people form the drying industry

and academia to make accurate estimations and get approximate dryer dimensions

(Gong and Mujumdar, 2008).


42

4.2.2 Drying simulation setup with Simprosys

The main reason for using Simprosys was to calculate the necessary air flow to

dry the bagasse coming from the dewatering mills. Figure 25 shows the layout of

the whole drying system as it was displayed in Simprosys. It represents the most

important elements of Figure 24 as well as a few smaller elements, like an air

filter and cyclone. The heater (H in Figure 25) represents the solar field and

exhaust steam heating system.

Figure 25: Simprosys model layout.

For this study the bagasse flow rate and moisture content was known, along with

the desired bagasse moisture content at the end of the dryer. Furthermore, the air

and bagasse inlet temperatures are known and outlet temperatures are assumed to

be the same, as suggested by Sosa-Arnao and Nebra (2009). Although the study of

Sosa-Arnao and Nebra (2009) uses a flash dryer, the outlet temperatures for

bagasse and air in this study is also assumed to be the same, this is not too an

unrealistic assumption, because of the longer residence time of the bagasse in a

rotary dryer. Another assumption that was made was that the air pre-heater would

be able to heat the air to 50 °C. Table 4 shows the input values which the program

had to use to calculate the necessary air flow.

As mentioned in the previous section, the program takes various heat losses and

inputs into account for the dryer. For this simulation it was assumed that there is

no heat input to the dryer except for the drying air. Furthermore, there is no

transport device for the bagasse inside the dryer, so there is not heat loss by the

transport device and it was assumed that the mechanical work done by the dryer is

only to improve the air-bagasse contact surface and that it did not have an

influence on the energy balance of Equation 4.1. Zabaniotou (2000) gives the heat

loss from a rotary dryer as 10 – 12 % of the total heat transferred, based on the

work of Keey (1994). So for this simulation it was assumed to be only 10 % of the

transferred heat, because of the relatively low temperatures compared to other

drying systems.


43

Table 4: Input values for Simprosys simulation.

4.2.3 Simulation results

Using all this information the program calculated that 9.49 MW is needed to heat

the air from the pre-heated 50 °C to the 120 °C for the dryer inlet. This is the

heating load that the solar system and exhaust steam heating system will have to

provide. The calculated flow rates are presented in Table 5. As mentioned in the

previous section, it is assumed that the ambient air can be pre-heated to 50 °C, this

should be possible as the air from the cyclone, which is used to heat it, enters the

pre-heater at 75 °C and calculated to exit the pre-heater at 56.67 °C.

After the heat and mass balances were solved, the program’s scoping function was

used to determine the size of the dryer needed. A gas velocity of 3 m/s was

chosen, as recommended by Kemp and David (2002), and the length/diameter

ratio was set to 5, which is in the range of effective ratios for rotary dryer design.

Simprosys recommends a dryer with a diameter of 7.8 m and a length of 39 m,

which is extremely large. This extreme size is due to the very high mass flow rate

of bagasse. In order to create a realistically sized drying system it was decided to

divide the bagasse flow into four equal streams, so that each dryer dries 10.9 ton/h

of bagasse. If the heat loss is adjusted accordingly, then all the flow rates listed in

Table 5 would be a quarter of what they were. Each rotary dryer will now have a

diameter of 3.9 m and a length of 19.5 m. The heat load for each dryer is now

2.37 MW.

Table 5: Simprosys outputs

Stream Temperature [°C] Flow rate [ton/h]

Absolute

humidity/moisture

content [kg/kg]

Heated air 120 479.169 0.016

Dry Bagasse 75 34.947 0.4

Air to cyclone 75 487.139 0.033

Wet exit air 56.67 487.139 0.033

Stream Temperature [°C] Flow rate

[ton/h]

Absolute

humidity/moisture

content [kg/kg]

Inlet air 25 - 0.016

Air to heater 50 - 0.016

Heated air 120 - 0.016

Wet bagasse 64.5 43.63 0.51

Dry Bagasse 75 - 0.4

Air to cyclone 75 - -


44

One advantage of having four dryers is that the boilers can still receive dry

bagasse if one of the rotary dryers fail or needs maintenance. The bagasse which

would normally go to the out-of-order dryer just needs to be mixed with the dried

bagasse before it enters the boiler as noted in the literature review. Furthermore,

by dividing the air stream into four, the necessary fans required can also be of a

more convenient size, more will be said about this in the results of pressure drop

calculations in Section 4.3.4.

4.3 Solar Field Modelling and Simulation

As mentioned previously, it was decided to use evacuated tube air collectors for

the solar system. Unfortunately, none of the well-established drying simulation

programs (like Polysun, T*Sol or SAM) allow for air collectors. However, there is

a program developed by Fraunhofer ISE, called Aircow(Welz, 2017), which

transforms the thermohydraulic collector parameters of an air collector into

coefficients of an equivalent liquid heating solar thermal collector. For Aircow to

do this it needs measured or simulated operational parameters and to find these a

model of an evacuated tube air collector was created in Matlab. This section will

discuss the Matlab model and the use of Aircow to determine the thermohydraulic

collector parameters which can then be used for the simulation of the solar field in

Solgain (Ilchmann et al., 2016).

4.3.1 Modelling of an evacuated tube air collector

It was decided to model a collector very similar to the Airwasol collector seen in

Figure 26, a brochure of the collector containing more technical specifications can

be found in Appendix A. This design was chosen because of its simplicity. To

model a collector it was decided to start with just one tube, see Figure 26.b,

because the heat transfer in each tube would be more or less the same.

(a)

(b)

Figure 26: (a) Airwasol air collector. (b) One of the evacuated tubes in the Airwasol air

collector (Siems, 2017).


45

To model the evacuated tube it was divided into little segments of 1 cm and it was

assumed that the glass and absorber temperatures are constant over these littles

segments in order to accurately calculate the thermal physical properties of air.

For each one these segments the heat transfer to the air inside and outside the tube

was determined. This was done through various calculations, using thermal

resistances as set out in Figure 27.

Figure 27: Schematic of heat transfer resistances in an evacuated tube.

Table 6 gives the characteristic of the evacuated tube as specified by Airwasol.

They also specified that the glass tube is made of Borosilicate 3.3 and that the

absorber tube is made from stainless steel. Appendix B shows how the

transmissivity of the Borosilicate was determined by taking the wavelengths of

the radiation into account following the method specified by Kalogirou (2009).

Table 6: Airwasol evacuated tube characteristics

Characteristic Unit Value

Length m 2

Glass outer diameter m 0.09

Glass inner diameter m 0.085

Absorber tube outer diameter m 0.0505

Absorber tube inner diameter m 0.0501

Absorber absorptance - 0.95

To calculate the glass temperature (Tg) the heat loss through natural convection

and radiation losses to ambient conditions were taken into account. For the natural

convection between the air and the glass an empirical correlation for the average

Nusselt number (Nu) over a horizontal cylinder was used, shown in Equation 4.6

as set out in Çengel and Ghajar (2015), who references Churchill and Chu’s 1975

paper. The work of Çengel and Ghajar (2015) is used for all the heat transfer


46

calculations of this section, unless stated otherwise. Appendix B gives all the

formula’s used to determine the necessary thermal physical properties of air.

𝑁𝑢 = (0.6 + 0.387𝑅𝑎𝐷

1/6

[1+(0.559/𝑃𝑟)9/16]8/27)2

(4.6)

Where the Raleigh number (Ra) was calculated as defined in Equation 4.7 and

must be smaller than 1012:

𝑅𝑎𝐷 = 9.81𝛽(𝑇𝑔−𝑇∞)𝐷𝑔

3

𝜈2 𝑃𝑟 (4.7)

Once Nu is calculated, the heat transfer coefficient is determined as well as the

thermal resistance of natural convection between the glass and the ambient air.

The thermal radiation resistance between the glass and sky must also be calculated

by first determining the heat transfer coefficient for thermal radiation between the

glass and sky, see Equation 4.8. σ represents the Stefan-Boltzmann constant

(5.67 x 10 -8 W/(m2K4)).

ℎ𝑟𝑎𝑑 = 𝜀𝑔 σ(𝑇𝑔2 + 𝑇𝑠𝑘𝑦

2)(𝑇𝑔 + 𝑇𝑠𝑘𝑦) (4.8)

A perfect vacuum was assumed between the glass tube and the steel absorber,

resulting in no convection losses between them. Therefore, it is only necessary to

calculate the thermal radiation resistance between the two. Equation 4.9 was

adapted from Çengel&Ghajar (2015), see Appendix B, in order to calculate the

thermal resistance (Rag) between the two tubes. Here εa and εg represent the

emissivity of the absorber and the glass, respectively.

𝑅𝑎𝑔 =

1

𝜀𝑎+

1−𝜀𝑔

𝜀𝑔(

𝑟𝑎𝑟𝑔

)

𝜎𝐴𝑎(𝑇𝑎2+𝑇𝑔

2)(𝑇𝑎+𝑇𝑔) (4.9)

Once all the thermal resistances were calculated the temperature of the glass cover

could be determined. This was done by using Equation 4.10 (see Appendix B for

derivation) where the radiation absorbed by the glass (Qrad,a) along with the

convection and radiation losses were taken into account to calculate the glass

temperature. As is evident in Equations 4.6 – 4.9, a glass temperature needs to be

guessed to determine some of the values needed for Equation 4.10. Therefore, to

accurately determine the glass temperature the guessed glass temperature needs to

be compared to the calculated glass temperature. If the error between the two is

not very small, the calculations need to be iterated until the difference is

negligible. For this study an error of 0.001 was deemed small enough.

𝑇𝑔 =𝑄𝑟𝑎𝑑,𝑎+

𝑇𝑎𝑅𝑎𝑔

+𝑇∞

𝑅𝑔∞+

𝑇𝑠𝑘𝑦

𝑅𝑔𝑠𝑘𝑦1

𝑅𝑎𝑔+

1

𝑅𝑔∞+

1

𝑅𝑔𝑠𝑘𝑦

(4.10)


47

The temperature of the absorber was calculated in a very similar way. The

radiation onto the absorber was calculated using the area of the absorber which

faces the sun and the amount of radiation that passes through the glass. The

Nusselt number for the forced internal convection between the absorber and the

air flowing inside of it was determined using Gnielinski’s equation, see

Equation 4.11. Gnielinski’s equation was used because of the expected high mass

flows and turbulent flow through the pipe due to the high air demand of the rotary

dryers. It was assumed that the turbulent flow in the tube is fully developed over

the entire length of the tube as the length/diameter ratio is larger than 10. Another

reason for the use of Gnielinski’s equation is its accuracy; an error of less than

10 % can be expected when using this formula according to Çengel&Ghajar

(2015).

𝑁𝑢 = (𝑓/8)(𝑅𝑒−1000)𝑃𝑟

1+12.7(𝑓/8)0.5(𝑃𝑟2/3−1)( 0.5 ≤𝑃𝑟 ≤2000

3×103<𝑅𝑒<5×106) (4.11)

Where the friction factor (f) is calculated using the first Petukhov equation

(Equation 4.12) as recommended by Çengel&Ghajar (2015).

𝑓 = (0.790 ln(𝑅𝑒) − 1.64)−2 (4.12)

The Nusselt number was again used to determine the heat transfer coefficient and

this was in turn used to calculate the thermal resistance to the internal forced

convection (Rair). The thermal resistance for radiation heat transfer between the

absorber and glass was already calculated to determine the glass temperature so

now the absorber temperature can be calculated as shown in Equation 4.13 (see

Appendix B for derivation). As with the calculation of the glass temperature, an

absorber temperature had to be assumed before the actual temperature could be

determined and the two temperatures needs to have an error smaller than 0.001.

𝑇𝑎 = 𝑄𝑟𝑎𝑑+

𝑇𝑔

𝑅𝑎𝑔+

𝑇𝑎𝑖𝑟𝑅𝑎𝑖𝑟

1

𝑅𝑎𝑔+

1

𝑅𝑎𝑖𝑟

(4.13)

The air temperature can be calculated using the absorber temperature and the

thermal resistance to the internal forced convection. The derivation of

Equation 4.14 can be found in Appendix B. For this calculation the inlet air

temperature (Tair,i) does not have to be guessed, because the air into the first

segment is pre-heated to a constant temperature and for the segments that follow,

the inlet temperature will be that of the previous segments’ output.

𝑇𝑎𝑖𝑟,𝑖+1 = 𝑇𝑎𝑖𝑟,𝑖 + 𝑇𝑎−𝑇𝑎𝑖𝑟

𝑅𝑎𝑖𝑟 �̇� 𝑐𝑝,𝑎𝑖𝑟 (4.14)


48

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 220

30

40

50

60

70

80

90

Tube Length [m]

Tem

pera

ture

[C

]

Air Temp

Absorber Temp

Glass Temp

4.3.2 Matlab simulation results

For this simulation the ambient air temperature was set to 25 °C, with a sky

temperature of 20 °C. The solar irradiance onto the tube was set to 1000 W/m2

and the air mass flow was 0.0108 kg/s. Figure 28 shows the simulation’s

calculated temperature rise across the evacuated tube for the glass, absorber and

air.

Unfortunately an experimental study could not be done to validate the

simulation’s results. There is, however, a study on evacuated tube air collectors

done in Canada by Paradis et al. (2015) where a similar model was created and

then validated with experimental results. In order to see how the current study’s

model holds up against the model of Paradis et al. (2015) their experimental

conditions were fed into this study’s Matlab model and the evacuated tubes were

set to the same size. By comparing the model like this, it ensures that the heat

transfer modelling was at least done correctly. Appendix C shows a further

comparison of the two models’ outlet air temperature and efficiency.

Figure 28: Matlab simulation results for the evacuated tube

For their experiment, they tried to reach a steady state operating condition by

providing constant volume flow. During the experiment most of the other

parameters like the ambient air and solar irradiance also stayed relatively constant.

As can be seen in Figure 29, the tube outlet temperature tends to 278 K when the

volume flow rate is at 28 m3/h. If the same flow rate and ambient air conditions

are fed into the current study’s model the outlet air temperature is calculated to be

277.87 K, which is only 0.05 % less than the value measured by Paradis et al.

(2015), showing that the model can be reliable. The size of the segments were

varied as well, to ensure that the model output does not rely on segment size; if


49

the segment sizes are set to 20 cm instead of 1 cm, the change in the output air

temperature is only 0.4 %.

Paradis et al. (2015) also incorporated wind speed into their model and

experimental results, but showed that it had little effect on the evacuated tube.

This was not a factor this model took into account, but it was not deemed

necessary due to the Paradis et al.’s (2015) results and because the Matlab model

showed that the collector was quite insensitive to changes in the glass

temperature. This can be explained by looking at Equation 4.13, here the thermal

resistance for radiation heat transfer between the absorber and glass is quite high,

minimizing the effect of the glass tube temperature.

Figure 29: Extract of Paradis et al. (2015) experimental results. (a) Ambient air temperature,

simulation output temperature and experimental output temperature as a function of time.

(b) Tilted irradiance, horizontal irradiance and reflected irradiance. (c) Wind speed during

testing. (d) Volume flow rate through evacuated tube.


50

4.3.3 System planning using Aircow

Simulated results from the Matlab model described above are given as an input

for Aircow. Aircow uses these inputs to calculate the air collector model

coefficients in order to optimise the system design. The coefficients can also be

converted into equivalent liquid heating collector coefficients of efficiency so that

it can be simulated using a conventional simulation tool. All of the information

regarding the program comes from the Aircow manual (Fraunhofer ISE, 2017).

The program’s optimisation function helps to determine the ideal mass flow, field

length and field width. The system planning function also calculates the solar

field’s total pressure drop and necessary fan power. One important thing to note is

that Aircow has only been validated for flat plate collectors with under-flown

absorbers and not evacuated tube collectors. This may lead to small errors,

however, looking at the program’s operations, it does not seem if it would make a

difference if a flat plate collector or evacuated tube collector is used.

Aircow considers the air collector’s efficiency as a map instead of a line as is

common with liquid heating collectors. This is because it uses both the reduced

temperature difference and the air mass flow rate as input values. Equation 4.15

shows the thermal efficiency model Aircow uses to determine the collector’s

conversion factor (η0), its linear heat loss coefficient (C1), its quadratic heat loss

coefficient (C2) and the mas flow dependent heat loss coefficient (Cm). The

collector efficiency (η), input-output temperature difference (∆T), solar irradiance

(Gt) and mass flow rate (�̇�) are all user inputs and in this study is determined

using the Matlab model desribed in the previous section.

𝜂 (∆𝑇

𝐺𝑡, �̇�) = (1 − 𝑒−𝐶𝑚 �̇�)(𝜂0 − 𝐶1 𝑇′ − 𝐶2 𝐺𝑡 𝑇′2) (4.15)

Where T’ is calculated by:

𝑇′ =(

𝑇𝑖𝑛+𝑇𝑜𝑢𝑡2

)−𝑇∞

𝐺𝑡 (4.16)

Aircow specifies that the simulated results, which were given as inputs, must

represent turbulent flow. Furthermore, it is important to give results from different

flow rates in order for the model to determine the various coefficients and

construct the efficiency map.

The program uses the calculated conversion factor and heat loss coefficients along

with the user defined solar radiation and collector tilt to determine the size of the

solar field that is needed to supply the needed amount of thermal power specified.

This is done by determining the energy the solar field can produce per m2 and

then dividing the design power output by this amount. The number of collectors in

a row is determined using the needed temperature rise and the mass flow rate of


51

the air through the row. The program can then use the calculated solar field area

and number of collectors per row to determine the amount of rows in the solar

field.

As mentioned before, Aircow can also determine the field’s pressure drop. In

order to do this it has a specific model which determines a linear (R1) and

quadratic (R2)pressure drop coefficients, as can be seen in Equation 4.17. It also

determines these coefficients from user input data regarding simulation results of

the pressure loss in one collector for a given mass flow.

∆𝑝𝑐𝑜𝑙𝑙(�̇�) = 𝑅1 �̇� + 𝑅2 �̇�2 (4.17)

In order to determine the collector’s pressure drop, it can be seen as a manifold

with parallel flow (refer back to Figure 26.a) and the equation for the pressure

drop in a manifold was calculated as specified by Bajura& Jones (1976). For this

study Equation 4.18 was used to determine the pressure loss over the evacuated

tube.

𝑝1−𝑝2

𝜌𝑎=

𝑣312

2(

𝐴31

𝐴32)

2

+𝑣31

2

2[𝐶𝑇𝐷 + (

𝑓 𝐿

𝐷) + 𝐶𝑇𝐶 (

𝐴31

𝐴32)] (4.18)

The various parameters of Equation 4.18 are displayed in Figure 30.a. The

equation makes allowance for a change of area in the lateral, but for this study the

area is constant, therefore the ratio A31/A32 = 1. The term’s CTD and CTC represent

the minor losses as the flow enters the lateral from the distribution header and

then exits the lateral into the combining header. The term fL/D accounts for the

frictional losses in the lateral, the frictional losses in the headers are deemed

negligible. This is because, when the collectors are placed in a row, the evacuated

tube exit of one collector would directly attach to the entrance of the following

tube, as shown in Figure 30.b, at the joining of the tubes there is just a small

support. This means that the total header length in the system is considerably

shorter than the total evacuated tube length in the system, therefore, the total

pressure loss in the headers are much smaller than the total pressure loss in the

tubes. The fact that the headers are larger than the tubes with the air flowing

through them at a lower speed also significantly reduces the pressure losses.

Aircow then goes on to determine the pressure drop over the entire system

(∆psys,before,after) for a given volume flow (�̇�). The pressure drop due to the

equipment in the system, before and after the solar field, are taken into account

and allows the program to determine the fan power (Pfan) needed for the given air

flow as shown in Equation 4.20. The pressure drop before and after the solar field

can be determined by using user specified flow resistances (R), see Equation 4.19.

The flow resistances of the field can be determined by calculating the estimated

pressure drop over the field for a given volume flow; basically just rearranging

Equation 4.19 so that R is on the left hand side.


52

The model also takes dynamic pressure (psys,dyn) into account, for when the air

flow velocity reduces due to flowing into a large volume, but for this study it was

not considered, since the air does not flow into a large volume, only the rotary

dryer. The pressure loss over the rotary dryer is accounted for in the pressure drop

after the solar field. Another function which will not be used is Aircow’s air

leakage model, this is because the Airwasol collector which this study tries to

model, does not have any leakage losses (Siems, 2017).

(a)

(b)

Figure 30: (a) Flow in a manifold’s lateral from Bajura& Jones (1976). (b) Airwasol

collectors connected in a row, with headers on the far left and far right and supports in

between (Siems, 2017).

∆𝑝𝑠𝑦𝑠𝑏𝑒𝑓𝑜𝑟𝑒,𝑎𝑓𝑡𝑒𝑟= 𝑅𝑠𝑦𝑠𝑏𝑒𝑓𝑜𝑟𝑒,𝑎𝑓𝑡𝑒𝑟

�̇�2 (4.19)

𝑃𝑓𝑎𝑛 =(∆𝑝𝑠𝑦𝑠𝑏𝑒𝑓𝑜𝑟𝑒

+∆𝑝𝑠𝑦𝑠𝑎𝑓𝑡𝑒𝑟+𝑝𝑠𝑦𝑠𝑑𝑦𝑛

+∆𝑝𝑐𝑜𝑙𝑙)�̇�

𝜂𝑓𝑎𝑛 (4.20)

The main goal of using Aircow is to convert the collector coefficients so that it

can be used in a simulation program for liquid heating collectors. Aircow

specifically mentions ScenoCalc as a simulation program that can be used, but the

converted coefficients are not limited to this software. To determine the

coefficients for a liquid heating collector, one operating mass flow is considered

for Equations 4.21-4.23. Appendix D shows how the mass flow can be optimized

usingAircow. Only one operating point is necessary for a collector without

leakages, if the collector experiences any air leakages, three operation points need

to be considered. The converted coefficients can now be used in Equation 4.24, as

is common for liquid heating collectors.

𝜂0 = 𝜂0𝑚𝑎𝑥(1 − 𝑒−𝑐𝑚 �̇�𝑟𝑜𝑤) (4.21)

𝐶1 = 𝐶1𝑚𝑎𝑥(1 − 𝑒−𝑐𝑚 �̇�𝑟𝑜𝑤) (4.22)


53

𝐶2 = 𝐶2𝑚𝑎𝑥(1 − 𝑒−𝑐𝑚 �̇�𝑟𝑜𝑤) (4.23)

𝜂 (∆𝑇

𝐺𝑡) = 𝜂0 − 𝐶1 𝑇′ − 𝐶2 𝑇′2 (4.24)

4.3.4 Aircow results

Simulation results from the Matlab model described previously was used as data

points for Aircow in order to determine the various heat loss coefficients and the

efficiency map. The mass flow, irradiance and inlet temperature for the Matlab

model was varied in order to give Aircow a wide range of data. Table 7 shows the

7 data points entered for Aircow. The mass flow through the collector was

determined by multiplying the mass flow through one tube, as specified in the

Matlab model, by 17, which is the amount of tubes in one Airwasol collector.

Table 7: Data points for Aircow input

The collector’s efficiency increases with an increase in mass flow, but decreases

for an increase in temperature as expected from air collectors. The last data point

does not follow the same pattern as the ones before it, this point was used to try

and simulate a very efficient point. However, the collector’s efficiency does

appear to be quite low. The main reason for this is the large glass tube diameter

compared to the absorber diameter. The efficiency was determined with

Equation 4.25. By looking at the equation, it is possible to see that the area of the

collector plays a very important role. If the absorber area was used instead of the

glass tube outside area the efficiency would be in the range of 0.7 – 0.8.

𝜂𝑐𝑜𝑙𝑙 =�̇� 𝑐𝑝 ∆𝑇

𝐺𝑡(𝐴𝑔

2)

(4.25)

Aircow determined the efficiency map and with that it also calculated the various

coefficients given in Table 8. To determine the pressure drop over the collector

Equation 4.18 was used, with the same mass flows as listed in Table 7. In

Irradiance

[W/m2]

Mass flow

[kg/h]

Inlet Temp

[°C]

Outlet Temp

[°C]

Collector

Efficiency

[%]

700 593.64 50 58.87 43.76

1100 593.64 50 64.08 44.20

700 1242.36 70 74.15 42.93

1100 1242.36 70 76.67 43.90

700 2282.76 100 102.13 40.68

1100 2282.76 100 103.51 42.55

1100 4284 25 27.04 46.16


54

Equation 4.18, CTD as well as CTC was set to 2 to account for the minor losses in

the threaded T-junctions where the evacuated tubes join the distributer and

collection headers (Çengel & Cimbala, 2010). Aircow gives R1 and R2 of

Equation 4.17 as -0.0263 Pa/kg/h and 0.000224 Pa/(kg/h)2, respectively.

Table 8: Evacuated tube quadratic model coefficients as determined by Aircow.

To optimise the mass flow and to determine the equivalent liquid heating collector

coefficients, a design point has to be set. The peak operation point was chosen for

these inputs. Therefore, the irradiance was taken as 1060 W/m2, which was the

highest global tilted irradiance measured during the crushing season, it was

measured on 31 October. The ambient temperature was set to 25 °C. The collector

tilt was set to 29 °, which is Durban’s latitude and the needed temperature rise

over the solar field needs to be 70 °C, with an inlet temperature of 50 °C. The

thermal power transferred to the air at the field outlet needs to be 9.49 MW, the

total amount needed to dry all of the bagasse. Using the above information

Aircow determined that there should be 5.88 collectors in a row and the solar field

should consist of 755.39 rows, which equates to a solar field size of 22 165.83 m2.

The flow resistances before and after the field, as mentioned in Equation 4.19,

was found to be:

𝑅𝑏𝑒𝑓𝑜𝑟𝑒 = 6.33 × 10−9𝑃𝑎

(𝑚3

ℎ)

2

𝑅𝑎𝑓𝑡𝑒𝑟 = 1.656 × 10−8𝑃𝑎

(𝑚3

ℎ)

2

For this study the flow optimisation for the net system power savings and for cost

savings was compared, that is options 2 and 4 set out in Appendix D. The price

per unit electrical auxiliary energy is given as 0.50 ZAR/kWh (Hess et al., 2016)

and the price of the energy replaced is 0.168 cent/kWh (see Appendix D for

calculation). For the cost optimisation Aircow suggests a mass flow rate of

510.9 kg/h per row and for the system savings optimisation a mass flow of

639.8 kg/h per row is suggested. The lowest flow rate that should be allowed in

the evacuated tubes is 550 kg/h, because this is the lowest flow rate where the

internal flow’s Reynolds number is still above 10 000. The Matlab model and

Quadratic model inputs Units Calculated values

Conversion factor (η0) - 0.4613

Linear heat loss coefficient (C1,max) W/(m2K) 0.440

Quadratic heat loss coefficient (C2,max) W/(m2K) 0.00859

Mass flow dependent heat loss coefficient (Cm) h/kg 0.007598


55

Aircow both need turbulent flow values, therefore the system power savings

optimisation flow rate will be used.

In reality there cannot be a decimal amount of collectors in a row. By following

the steps set out in Appendix D, the system is recalculated with 6 collectors in a

row, but with 749 rows, so that the power at the field outlet will still be the same.

However, the outlet temperature will be 121.3 °C. This temperature is not

dangerous for the drying system, because it is not high enough to cause a fire in

the system. Furthermore, the temperature of the air which reaches the drying

system will be lower than the air temperature directly at the field outlet, due to

heat losses in the piping system. The instantaneous efficiency of this point

according to Aircow is 40.3 %, which is quite low. This is most probably due to

the high temperatures and low flow rate in the solar field, with both factors

negatively impacting the efficiency.

For the optimised system, Aircow calculates that the pressure drop over the solar

field will be 449.2 Pa and 6 105.5 Pa over the rest of the system. The pressure

drop is quite high due to the rotary dryer, which accounts for half of the drop. So

in total the system has a pressure drop of 6 554.7 Pa with a volume flow rate of

444 438.9 m3/h. If a fan with an electric to hydraulic efficiency of 57.2 % (an

Aircow default value) is used, the fan will require 1.41 MWel. This could result in

a very large fan being used, therefore it is suggested that, like the drying system,

the solar system is also split into four sections. Each solar section will provide the

necessary heated air to the rotary dryer it is connected to and a smaller fan of a

more convenient size can then be installed into each section.

The flow rate in each collector row will stay the same, as well as the number of

collectors in each row, since the same temperature lift is still needed. The number

of rows for each section will just be a quarter of the rows specified above. The

volume flow rate for each section will also be a quarter of what is specified above.

It is important that the mass flow rate through the solar system is the same mass

flow rate that is needed in the drying system, because it would lead do efficiency

losses if extra air had to be sucked in from the ambient or if excess air had to be

released. The pressure drop across each section is assumed to be the same,

therefore, the fans’ will still have to produce the necessary pressure lift, but

because the volume flow rate in each system is a quarter of what was mentioned

above, each fan can be much smaller.

Now that a specific operation point has been identified, Aircow can calculate the

coefficients that can be used in a liquid collector simulation, as shown in Table 9.


56

Table 9: Converted coefficients for liquid collector simulation.

4.4 Solar Field Simulation

For this study Solgain (Ilchmann et al., 2016) was chosen as simulation program

for the evacuated tube collector system rather than ScenoCalc. This was done,

because ScenoCalc does not take the whole solar field into account, only the

power per m2. Furthermore, it does not allow the user to specify temperatures

above 100 °C, something which is necessary for this study. Solgain is an open

source tool which can be used to simulate the entire solar system and it can be

edited to function with temperatures above 100 °C. Solgain was created to

accurately and quickly determine the solar gains of a system without too much

effort from the user (Ilchmann et al., 2016).

4.4.1 Simulation setup with Solgain

In order to simulate a solar thermal system the user should give inputs regarding

the climate, location and process. For the climate data, hourly GHI values, DNI

values and the ambient air temperature should be given as inputs. For the location

input, the latitude and hemisphere of the system must be given. The climate and

location data is used to determine the GTI (named Gt in Solgain), which is the

total irradiance that falls on the collector’s surface and was calculated using

Equation 4.25.

𝐺𝑡 = 𝐺𝑏𝑡 + 𝐺𝑠𝑡 + 𝐺𝑟𝑡 (4.25)

Gbt represents the direct irradiance on the tilted collector aperture, Gst is the

diffuse irradiance on the tilted collector aperture and Grt is the reflected irradiance

onto the tilted collector aperture. These terms were calculated using

Equation 4.26- 4.28.

𝐺𝑏𝑡 = 𝐺𝑏𝑛 × cos𝜃 (4.26)

Where θ is the incidence angle, calculated for each hour and Gbn is the DNI.

𝐺𝑟𝑡 = 𝐺 (𝜌𝑔𝑟𝑑

2) (1 − cos𝛽𝑐) (4.27)

Where G is the GHI, ρgrd is the ground reflectivity and βc is the collector tilt.

Coefficients Units Converted values

Conversion factor (η0) - 0.4577

Linear heat loss coefficient (C1) W/(m2K) 0.437

Quadratic heat loss coefficient (C2) W/(m2K) 0.008524

Mass flow per row per row gross area kg/(sm2) 0.006

Mass flow per row per collector gross area kg/(sm2) 0.036


57

𝐺𝑠𝑡 = 𝐺𝑑 ×1+cos𝛽𝑐

2 (4.28)

Where Gd is the diffuse horizontal irradiance, which can be calculated by

subtracting the direct horizontal irradiance (Gb) from the GHI. Gbcan be

calculated as specified in Equation 4.29, where θs represents the zenith angle of

the sun.

𝐺𝑏 = 𝐺𝑏𝑛 × cos𝜃𝑠 (4.29)

After the program calculated the various irradiances that falls on the collector’s

aperture it can determine the specific collector gains (𝑄𝑔𝑎𝑖𝑛). This is done by

using the coefficients which were originally specified in Equation 4.24 and given

in Table 9. The incidence angle modifiers for direct (Kb) and diffuse (Kd)

irradiance are also used to determine the collector gain, as can be seen in Equation

4.30. Solgain already has specific collector coefficients for a flat plate collector

and for an evacuated tube collector, so for this study, the coding was changed so

that it uses the coefficients in Table 9.

𝑄𝑔𝑎𝑖𝑛 = 𝜂0[𝐾𝑏 𝐺𝑏𝑡 + 𝐾𝑑 𝐺𝑠𝑡 + 𝐾𝑑 𝐺𝑟𝑡] − 𝐶1(𝑇𝑓 − 𝑇𝑎) + 𝐶2(𝑇𝑓 − 𝑇𝑎)2

(4.30)

Solgain goes on to use the specific collector gains to calculate the field outlet

temperature. For this calculation Equation 4.31 is used. Solgain does not allow the

user to specify that more than one collector in a row is needed, it only calculates

the temperature rise over one collector and then assumes that this is the needed

outlet temperature of the field. So in order to simulate a field where there are

multiple collectors in a row, the coding needed to be changed. A loop was used

which took the first calculated outlet temperature and then used it as the inlet

temperature for the next collector. The outlet temperature was then again

calculated using Equation 4.31. The loop runs for the amount of collectors there

are in each row.

𝑇𝑜𝑢𝑡 =�̇� 𝑐𝑝𝑐 𝑇𝑖𝑛−

𝑐𝑒𝑓𝑓

2∆𝑡𝑇𝑖𝑛+𝑄𝑔𝑎𝑖𝑛+

𝑐𝑒𝑓𝑓

∆𝑡𝑇𝑓(𝑡−∆𝑡)

�̇� 𝑐𝑝𝑐+𝑐𝑒𝑓𝑓

2∆𝑡

(4.31)

In this equation the cp,c represents the average heat capacity, while ceff represents

the effective collector heat capacity (kJ/m2K). For the average heat capacity value,

the heat capacity of air between 50 °C and 120 °C was embedded into the

program, while the effective collector heat capacity was kept the same, because it

is related to the aperture, rather than the heat transfer fluid. ∆t represents the width

of the time steps Solgain work with, which is one hour, since hourly data is used

as input. For Equation 4.31, �̇� represents the specific mass flow through the

collector field. Tfis the mean collector temperature and is dependent on the outlet


58

temperature, therefore, Equations 4.30 and 4.31 along with Tf will have to be

iterated, where the error between outlet temperatures should be smaller than the

user defined value. When the outlet temperature is finally determined, the total

thermal energy delivered by the solar field can be determined using

Equation 4.32.

𝑄𝑐𝑜𝑙𝑙 = �̇�𝐶 𝑐𝑝𝑐(𝑇𝑜𝑢𝑡 − 𝑇𝑖𝑛) (4.32)

Solgain assumes that there is a heat exchanger between the solar field and the

process to which it delivers the thermal energy, as can be seen in Figure 31. The

integration point for this study does not include a heat exchanger and to

effectively remove the heat exchanger its efficiency was just set to 100 %, so that

the heat delivered by the solar field equals the heat that goes to the process.

Solgain also allows the user to define the distance between the solar field and the

process. The program then calculates the amount of heat loss across this distance

to and from the process.

Figure 31: Solgain system layout (Ilchmann et al., 2016).

In order to determine how the system performs, Solgain completes three

calculations to describe the system’s characteristics. These calculations are that of

the solar fraction (F), degree of utilization (ηu) and the annual system

efficiency (ηs). The solar fraction is the percentage of thermal energy needed by

the process (Qneed) which the solar system can provide throughout the year (Qp).

The degree of utilization is the percentage of the energy provided by the solar

field (Qcoll) which the process can actually use. The annual system efficiency

shows how much of the sun’s energy (Qsun) the solar field could use and send to

the thermal process. The equations for the three characteristics are shown in

Equations 4.33 - 4.35.

𝐹 =𝑄𝑝

𝑄𝑛𝑒𝑒𝑑 (4.33)


59

𝜂𝑢 =𝑄𝑝

𝑄𝑐𝑜𝑙𝑙 (4.34)

𝜂𝑠 =𝑄𝑝

𝑄𝑠𝑢𝑛 (4.35)

4.4.2 Simulation results

The same solar resource data from Solargis©, referred to in the literature study, is

used for this simulation. As mentioned before, Solgain only needs the hourly DNI,

GHI and ambient temperature. Most of the other inputs are summarised in

Table 10.

Table 10: Inputs for Solgain

Input Unit Value/decision

Heat demand type - Pre-heating

Process feed temperature °C 120

Process Return Temperature °C 50

Heat capacity of process medium kJ/(kgK) 1.009

Energy consumption of the process kW 9492.251

Collector tilt ° 29

Collector orientation - North

Collector area m2 22 425.06

The ground reflection was set to 0.2, the default value in Solgain. The maximum

difference for the output temperature iteration is set to 0.01, again the default

value in Solgain. The distance between the collector and process is set as 20 m,

this is to ensure there is enough space for the exhaust steam heater, rotary dryer

drive system and the bagasse feeding system. The maximum collector power

output is set to 427.31 W/m2, this value is determined by dividing the maximum

power output from the solar field with the collector area calculated by Aircow.

For this study the crushing season is taken from 1 April to 20 November, so the

heat demand profile in Solgain was set that the system only operates between

weeks 14 – 47 in the year. It was assumed that the drying system will have the

same overall time efficiency as the rest of the sugar mill, therefore the load was

set to 80.83 %.

Table 11 gives the results of the system characteristics as it was calculated by

Solgain. Two types of results are given, one for the actual system, which would

only operate during the crushing season and one for a system which would

operate throughout the whole year. The solar fraction initially seems to be quite

low, but one has to remember that the drying system runs day and night and the

solar system is only able to produce thermal energy throughout the day. On top of

this, the solar field only comes near to producing the total needed amount of


60

thermal energy for drying close to solar noon for about 2-3 hours each day, as it is

essentially a pre-heating system. The solar fraction for the annual system is a bit

higher than for the crushing season, this is because the sun can provide more

power during the summer months, which is outside of the crushing season.

Table 11: Simulation characteristics as calculated by Solgain

The degree of utilization during the crushing season is quite low, this is because

the drying system cannot use the thermal energy produced by the solar field

outside of the crushing season. The degree of utilization for the annual system in

turn is quite high, the reason why it is not 100 %, is because of the piping losses

from the solar field to the process.

Figure 32 shows the relationship of the solar radiation available and the thermal

energy that the field can deliver at design conditions. This is a representation of

the system’s efficiency, since it compares the radiation flux onto the solar field

with the energy delivered to the drying system. The annual system efficiency in

both cases are quite low, usually it is in the range of 40 – 60 % (Ilchmann et al.,

2016). This is due to the low collector efficiency, especially the fact that the glass

tubes are considerably bigger than the absorber tubes. This results in an aperture

area which is much larger than the area that could actually use the irradiation.

Therefore, by looking at Equation 4.35, the term Qsun is calculated using the

aperture area, while Qp depends on the absorber area.

Figure 32: Comparison of available solar radiation and thermal energy delivered to process.

0

5

10

15

20

25

1 3 5 7 9 11 13 15 17 19 21 23

The

rmal

En

erg

y [M

W]

Hour

Radiation Flux

Solar ThermalEnergyDelivered

Characteristic Crushing Season [%] Annually [%]

Solar fraction 15.93 16.61

Degree of utilization 70.44 95.49

Annual system efficiency 12.88 19.46


61

Figure 33 shows the thermal energy delivered to the drying system during the

crushing season. The drying system needs 9.49 MW when running at full load, the

highest amount of thermal energy the solar system can deliver is 9.48 MW, just

below the maximum needed energy. This is due to the fact that the peak

conditions were used to size the field, as the solar system only serves as a type of

pre-heater, it would be unnecessary and wasteful if the air is heated up too much

and producing excess thermal energy. Now none of the solar thermal energy

gained will go to waste, everything can be used by the drying system. An excess

of solar thermal energy can lead to lower system efficiencies, because some of the

energy will have to be dumped before it heads to the drying system.

Figure 33: Simulated solar thermal energy delivered to the drying system.

4.5 Effect on Sugar Mill

By using the results from the drying model and Solgain simulation, the amount of

bagasse that is saved or the reduction in coal usage can be determined.

Equation 4.36 can be used to determine the amount of dry bagasse (�̇�𝑏𝑎𝑔) that

needs to be used to produce the same amount of live steam (�̇�𝑙𝑠) than the wet

bagasse.

�̇�𝑏𝑎𝑔 = (�̇�𝑙𝑠ℎ𝑙𝑠)−(�̇�𝑏𝑓𝑤ℎ𝑏𝑓𝑤)

𝜂𝑁𝐶𝑉 𝑁𝐶𝑉 (4.36)

The boiler’s efficiency (𝜂𝑁𝐶𝑉) based on the NCV of bagasse was calculated to be

87.48 % when the boiler operates with wet bagasse. This is very close to the value

simulated by Laubscher (2017) when looking at Figure 15. His simulation roughly

shows a 2.45 % increase in boiler efficiency with the use of 40 % moisture

0

1

2

3

4

5

6

7

8

9

10

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Sola

r th

erm

al e

ne

rgy

[MW

]

Month


62

content bagasse. Assuming that the generic boiler will respond in the same way

and by using Equation 4.36 it was calculated that only 32.64 ton/h of the dried

bagasse is needed to produce the necessary live steam. It is, however, important to

keep in mind that the 40 % moisture content bagasse is lighter than the wet

bagasse, due to the reduction in moisture. Therefore, when looking at the actual

amount fibres used, 21.38 ton/h is needed of the wet bagasse and 19.58 ton/h is

needed of the dry bagasse.

The amount of exhaust steam (�̇�𝑒𝑥ℎ𝑠) required to deliver all of the needed

thermal energy (𝑄𝑡ℎ,𝑑𝑟𝑦) to the drying system can be calculated using

Equation 4.37. Where ℎ𝑒𝑥ℎ𝑠 and ℎ𝑐𝑜𝑛𝑑 represent the enthalpy of the exhaust steam

and condensate respectively. According to Rein (2007) one ton of live steam can

be let-down to 1.2 tons of exhaust steam, by taking this into account and using

Equation 4.36 and 4.37, it can be calculated that an extra 0.72 ton/h of the dry

fibres are needed to generate the needed exhaust steam. Therefore, the total

amount of dry fibres needed when using 40 % moisture content bagasse is

20.3 ton/h.

�̇�𝑒𝑥ℎ𝑠 =𝑄𝑡ℎ,𝑑𝑟𝑦

ℎ𝑒𝑥ℎ𝑠−ℎ𝑐𝑜𝑛𝑑 (4.37)

The solar field can help to save more bagasse, since it decreases the amount of

exhaust steam needed for the drying system. The previous section mentions that

the solar system can deliver 15.93 % of the necessary heat for the drying system.

This relates to 6 833.16 MWhth energy delivered. By multiplying this value with

3600 to get the energy delivered over the given period and plugging this value

into Equation 4.37, the amount of exhaust steam it saved can be determined. By

using this method it is calculated that the solar system saves 11 164.94 ton of

exhaust steam throughout the crushing season, this relates to 3 140.62 ton of dry

bagasse or 1884.37 ton dry fibre. This leads to a further 2.05 % reduction in

bagasse usage, therefore by integrating the bagasse dryer with the solar system the

bagasse usage can be reduced with 7.05 %.

To give a better idea of the value added by the solar system the amount of bagasse

saved can be expressed in tons of coal. This is done, because bagasse does not

have a specific monetary value, while coal does. The amount of coal saved can be

determined by comparing the calorific values of the two fuels. According to Smith

et al. (2016), 1 ton of coal is equal to 4 tons of bagasse. Based on NCV, 1 ton the

dried bagasse is equal to 1.3 tons of the 51 % moisture content bagasse. Therefore

1 ton of coal is equal to 3.08 tons of dried bagasse and it can be calculated that the

solar system will save 1020.38 tons of coal.


63

5 Solar Live Steam Generation

This section will explain how the proposed live steam generation system will

work using CSP collector technology. It will briefly discuss how the boiler and

solar system will work together, as well as how more electricity can be generated

in South African sugar mills. The section goes on to explain how the solar system

was simulated and the simulated results will be presented and discussed.

5.1 Integration Point

The integration of solar live steam into the generic system shown in Figure 3 in

Section 2.2 will help to ease the boilers’ load, saving bagasse and/or coal. This

study assumes that the generic sugar mill has three boilers which run at part load

to supply the necessary steam to the mill. They are sized similar to what is

described in the study of Reid and Rein (1983) where two boilers running at

110 % load would be able to supply the necessary steam (Reid and Rein, 1983).

Furthermore, each individual boiler can only be turned down to 50 % of its

nominal load capacity, resulting that the boiler house as a whole can only turn

down to 68 % of its load if all three boilers are running. This configuration

severely limits the size of the solar system and the impact it can make on the sugar

mill.

Although a solar thermal system would make it possible to shut down one or more

of the boilers during daytime, the boiler would have to be started again during

night time or hours of low DNI. Sugar mills refrain from shutting down boilers

and starting it up again on a regular bases as this may cause various problems in

the boiler (Foxon, 2017). An extremely large solar thermal storage system would

be needed if it was to supply energy to the mill throughout the whole night.

The proposed solar system would consist of parabolic trough collectors using

thermal oil. This technology was chosen since it is the most mature solar CSP

technology on the market (Calamateo and Zhou, 2015) and South Africa has

various operational parabolic trough plants like Bokpoort CSP and KaXu Solar

One.

The fact that there are large parabolic trough plants in South Africa means that the

necessary practical knowledge to implement and operate such a system is

available within the country. Although central receiver systems are becoming

more popular around the world, there is no need for its high operational

temperature which is one of its key advantages. Furthermore, the high humidity in

the sugar milling region would significantly reduce the efficiency of a central

receiver system (Cardemil et al., 2013). Linear Fresnel systems were deemed too

immature, considering that there is only one small system installed in South

Africa. Immature technology can make it difficult to secure funding as financiers

want to have certainty over technology capabilities (Peterseim et al., 2013).


64

Steamboilers

Let downvalve

G

Exhaust steamCollector

Condensate

Live steam 390 °C | 31 bar | 96.76 t/h

120.2 °C | 2 bar90.85 t/h

Make-upwater

121 °C | 2 bar | 96.76 t/h

Feed watertank

Kettle-typeheat

exchanger

Boilerfeed water

113 °C | 2 bar96.96 t/h

Solar thermal system Integration Conventional system

Bagasse

Coal

25 °C | 1 bar6.11 t/h

Flue gasClear juicepre-heater

1st effectevaporator

Sugardrying

BPSTPrimemovers

Pressure valve

Circulation pump

Feedpump

Figure 34 shows how a parabolic trough solar system could be integrated into a

conventional sugar mill. One problem with using thermal oil is that it has a

maximum operating temperature of 393 °C (Heller, 2013). The oil will have to

pass through a heat exchanger to create steam, so the maximum temperature

which could be delivered to the sugar mills is 385 °C (Peterseim et al., 2014),

while the steam used in the mill is 390 °C. However, according to the SMRI this

will not be a problem, the temperature difference should just not be more than

10 °C (Foxon, 2017). The steam pressure is a more important parameter that

needs to be adhered to. This is the reason for the pressure valve in the integration

setup in Figure 34.

Figure 34: Solar live steam integration into a generic sugar mill, adapted from Hess et al.

(2017).

The required steam temperature is another reason why thermal storage was not

considered. As mentioned in Section 3.1, molten salt is usually used as storage

medium, therefore, a heat exchanger is needed between the oil and molten salt.

This would result in a temperature drop when the tank is being charged and

another temperature drop when the tank is being discharged, creating steam at a

temperature closer to 370 °C. Oil can also be used as storage medium, but this

would result in rather large storage facilities and high costs (Liu, et al., 2016).

As mentioned before, conventional sugar mills do not operate outside of the

crushing season. Therefore, the solar system will also not be able to function

outside of the crushing season. This is because there are no operations that can use

25 °C | 2 bar 6.11 t/h


65

Let downvalve

Solar live steam

Steam from boiler

CEST

To condenser

G

Exhaust steam to processes

Prime movers

the live steam generated by the solar system. If extra electricity was to be

generated, then there are no processes that can use and condense the exhaust

steam from the BPST. The configuration above, from here on referred to as

Configuration 1, will also not allow for extra electricity production during the

crushing season, because there are no processes that can use the extra exhaust

steam from the BPST.

For a sugar mill with a back-end refinery the solar field will unfortunately still not

be able to operate outside of the crushing season, because the steam consumption

of the refinery is so small compared to the mill, that only one boiler running at its

minimum load would be able to supply the necessary steam. The solar live steam

could also not be used generate extra electricity since there are no extra processes

that can use and condense the extra exhaust steam.

The fact that the solar system is out of use during large parts of the year will lead

to low annual efficiencies and a relatively expensive system. An option that would

allow the solar system to function outside of the crushing season as well will be

discussed in the following section.

5.2 Increasing Electricity Production

The previous section identified that the BPST indirectly stands in the way of extra

electricity production and whole year operation of the solar system. Therefore,

one option is to replace the BPST with a CEST and a condenser. Figure 35 shows

such a configuration, which would from here on be referred to as Configuration 2.

Figure 35: Schematic of Configuration 2

During the crushing season the steam passing through the CEST would be

extracted at the right conditions to supply the necessary exhaust steam. This

configuration will also eliminate the need for a let-down valve. This is because the


66

Let downvalve

Solar live steam

Steam from boiler

CEST

To condenser

G

Exhaust steam to processes

GBPSTPrime

movers

new CEST can be designed to handle additional live steam and therefore

additional exhaust steam can be extracted as well. The steam which would have

normally been let-down can now go to the CEST, creating extra electricity.

The amount of steam that passes through the new CEST is thus the amount of

steam that would have gone through the old BPST plus the steam that would have

gone through the let-down valve. Newer CEST technology is more efficient than

the old BPST used in South African sugar mills (Ensinas et al., 2007; Foxon,

2017) which will lead to a large increase in electricity generated.

The solar system would ease the boilers load during the crushing season and

would now be able to function on its own outside of the crushing season. This is

because the CEST would allow for the steam to be exhausted to below

atmospheric pressure before it heads to the condenser. This configuration can,

therefore, save bagasse or coal during the crushing season and create extra

electricity throughout the whole year.

Another option that will allow the solar system to be in use outside of the crushing

season is to implement the setup shown in Figure 36, hereon referred to as

Configuration 3. In this configuration a CEST is placed in parallel with the

existing BPST. This configuration would also need a condenser as well as a new

turbo alternator, making it more expensive than the previous configurations. It is a

similar configuration to what is used in various international sugar mills (Bhatt,

2014; Burin, et al., 2016).

Figure 36: Schematic of Configuration 3

For this configuration all the steam that is produced by the solar system

throughout the year can pass through the CEST, generating extra electricity. This

configuration would also allow the steam that would normally pass through the

let-down valve to go to the CEST, creating even more electricity during the

crushing season as is the case with Configuration 2. The BPST and boiler would


67

just go on to function as they did and the let-down valve could stay in place in

case of extra exhaust steam shortages or any other irregularities.

Sugar mills with back-end refineries will also be able to benefit from

Configurations 2 and 3. They will function the same way as the normal mills

described above during the crushing season. Outside of the crushing season, the

refinery part of the mill will still be operating, as mentioned before, one boiler

running at its minimum load would be able to supply the necessary steam.

Unfortunately the solar system cannot help to ease the boilers load, since it cannot

be turned down any further and the sugar mills want to avoid shutting down the

boiler and starting it up again on a regular basis. The solar system can, however,

provide extra steam to increase the power generation outside of the crushing

season for the mill with the refinery as well.

5.3 System Advisor Model Simulation Setup

The System Advisor Model (SAM) (National Renewable Energy Laboratory,

2017) is a simulation program developed for the renewable energy industry. It is

used in this study to make performance predictions of the proposed parabolic

trough system. SAM’s Process heat parabolic trough performance model was

used to simulate the solar fields, one for a sugar mill without a refinery and one

for a mill with a refinery

The Process heat parabolic trough performance model is very similar to the more

commonly used CSP parabolic trough (physical) model which is used for power

generation systems. The process heat model, however, does not take the power

block into account nor solar thermal storage. The design point is also determined

slightly differently with the process heat model, as it has a System Design setup.

Here the user can define the design point DNI, the target solar multiple, the heat

transfer fluid operating temperatures and the thermal power it needs to deliver.

This is basically a combination of the usual parabolic trough model’s Solar Field

and Power Cycle setups.

SAM calculates the aperture area of the solar field using the information from the

System Design setup, the loop optical efficiency and total loop conversion

efficiency which SAM determines using characteristics of the solar collector. The

model does not give the user the option to define the field aperture as with the

usual model. For the simulations done for this study, solar noon on the vernal

equinox (20 September) was chosen as the design point, resulting in a design

point DNI of 790 W/m2 being used.

Therminol VP-1 was chosen as heat transfer fluid and it is set to operate with a

loop inlet temperature of 290 °C and a loop outlet temperature of 393 °C. There

are four solar collector assemblies in each loop of the solar field which will ensure

that the heat transfer fluid is heated up to the required maximum temperature.


68

FE GmbH (formerly Flabeg GmbH) Ultimate Trough collectors along with Schott

PTR 70 receivers were used for the simulations. Both the collector and receiver

are available in SAM’s library, along with their default efficiency values, which is

what was used during this study. This is very similar to the solar collector

assemblies which were used at Bokpoort CSP near Groblershoop in South Africa

(National Renewable Energy Laboratory, 2017).

The main difference being that the Bokpoort CSP collectors are not as wide as the

Ultimate Trough collectors. This means that there will be less loops in the

simulated solar field compared to when the collectors of Bokpoort CSP would be

used, however, the impact in performance should be minimal as the optical

efficiencies stay the same. The decision to use the above mentioned collector

technology was based on the success Bokpoort CSP has had and the fact that this

means that there is an assembly plant operational in South Africa.

Equation 5.1 was used to determine the maximum flow rate of the heat transfer

fluid in each loop, to ensure that the solar system does not provide too much

thermal energy to the sugar mill. In the equation ṁHTF is the maximum loop flow

rate, QPTis the maximum thermal power the solar field needs to deliver, NLis the

number of loops in the solar field, cp,HTFis the specific heat of the heat transfer

fluid and T is the temperature difference of the heat transfer fluid at the inlet and

outlet of the solar field.

�̇�𝐻𝑇𝐹 = 𝑄𝑃𝑇

𝑁𝐿 𝑐𝑝,𝐻𝑇𝐹 ∆𝑇 (5.1)

The row spacing is set to 3 times the width of the collector as suggested by

Gunther et al. (2013), this should minimize shading without allowing the solar

field to become too large. The solar field was set so that is has a tracking axis in

the north-south direction. The length of the piping between the solar field and the

process was set to 100 m, quite a large distance, but as it is still uncertain how the

solar field would be situated in relation to the mill, it was decided to rather be

conservative and give a large value. This is conservative as SAM uses this

distance to model heat losses between the solar field and processes, therefore, the

longer the distance, the higher the heat loss.

Table 12 shows more of the solar field parameters used for the simulations. The

solar field size, actual solar multiple and number of loops are all automatically

calculated by SAM.


69

Table 12: Solar field parameters for the two SAM simulations.

5.4 System Advisor Model Simulation Results

Table 13 gives a summary of the simulation results as well as some key

performance factors for the solar system. The simulations show that the capacity

factors of the two systems are quite low. The capacity factor is the percentage of

energy the solar field creates compared to the energy the installation can create if

it was to run continuously. In most CSP systems without storage the capacity

factors range from 20-25 % (IRENA, 2012). The reason the systems’ capacity

factors are so low is because of the low annual DNI received, most CSP plants

receive almost double the DNI per year.

Table 13: SAM simulation results

The solar fields’ efficiencies compare well to what was found in literature (Giostri

et al., 2012) and incorporates the optical, thermal and piping efficiencies. The

effect of the piping length from the solar field to the heat exchanger does not seem

to make a major influence, the annual solar gains increases or decreases with less

than 1 % if the length is reduced or increased by 100 m.

The solar system designed for the normal sugar mill has higher annual solar field

efficiency, but a lower capacity factor compared to the solar system designed for

the mill with the refinery. This is due to its lower solar multiple. The solar

multiples are not exactly 1 due to the sizes of the collectors, they do not exactly fit

the theoretical solar field size, and therefore, the actual solar multiple is a bit more

than what is needed.

Parameter Unit Sugar Mill Sugar Mill with

Refinery

Maximum Solar Heat Production MW 22.93 30.23

Solar Field Size m2 41 280 55 040

Actual Solar Multiple - 1.08 1.09

Number of Loops - 6 8

Maximum Flow per Loop kg/s 15.22 15.05

Row Spacing m 24 24

System

Energy

Generated by

Solar Field

[MWh/a]

Capacity Factor Average

Solar Field

Efficiency

[%]

Crushing Season

[%]

Outside of

Crushing

Season [%]

Sugar Mill 27 549 13.0 14.8 49.4

Sugar Mill with Refinery 36 525 13.1 14.9 49.1


70

Figure 37 shows the solar field’s annual output for the normal sugar mill

simulation. From this image one can clearly see that there are quite a lot of days

with very low DNI, resulting in low to almost no thermal power being produced

by the system. The low DNI are due to cloudy and rainy days, September –

November especially hve many of these days due to the rainy season. Despite this,

the capacity factor outside of the crushing season is still higher than during the

crushing season. This is due to the fact that on the days that the sun actually

shines, the system can run at full capacity for almost the entire day.

The system performs worse during the winter months. The lower performance is

due to a combination of cosine losses and the incidence angle modifier. The fact

that the solar field does not perform very well during the winter does not bode

well for Configuration 1, as winter covers most of the crushing season and it is the

only time when it will operate.

Figure 37: Simulated annual solar field output

In order to counter the negative effects of the cosine losses and incidence angle

modifier during the winter months, it was decided to simulate a solar field with an

east-west tracking axis. As mentioned in Section 3.1.1, this can reduce the

difference in energy yield between the summer and winter months. This tracking

option also needs less power to track the sun, as it does not follow the sun on its

east-west path each day, but rather its north-south path as it changes throughout

the year.

However, the annual energy yield is expected to be lower than for the north-south

tracking axis system, because the east-west tracking axis’s performance during the

day is quite uneven due to large incidence angles close to sunrise and sunset.

Appendix E shows the output of the two different tracking systems for specific

days of the year to illustrate this.

0

5

10

15

20

25


Sola

r Th

erm

al E

ne

rgy

[MW

]


71

The simulated results of Figure 38 shows that the system can reach the maximum

thermal output during the winter, seemingly performing better than the north-

south tracking axis system. However, when looking at Table 14, one can see that

the systems perform better during the crushing season, but considerably worse

outside of the crushing season. This is because there are so may cloudy and rainy

days outside of the crushing season, and unlike the north-south tracking system,

the east-west tracking system does not operate at full load for long periods of the

day, therefore, it cannot take advantage of the higher solar resource during the

summer months as it can only reach its maximum for a short period during solar

noon.

Figure 38: Simulated annual solar field output for an east-west tracking axis system

The thermal output during the crushing season is calculated to be 5.2 % higher for

the east-west tracking system compared to the north-south tracking system. If the

performance of the whole year is considered, the east-west tracking system

produces 7 % less thermal power than the north-south tracking system. It would,

therefore, make sense to use the north-south tracking axis system for

Configuration 2 and 3, and then use the east-west tracking axis system for

Configuration 1.

Table 14: SAM simulation results for an east-west tracking system.

0

5

10

15

20

25


Sola

r Th

erm

al E

ne

rgy

[MW

]

System

Energy Generated

by Solar Field

[MWh/a]

Capacity Factor Average

Solar Field

Efficiency

[%]

Crushing

Season [%]

Outside of

Crushing

Season [%]

Sugar Mill 25 635 13.7 11.0 46

Sugar Mill with Refinery 33 943.5 13.8 11.1 45.6


72

5.5 Effect on Sugar Mill

This subsection will discuss the effect the integration point will have on the sugar

mill and how each of the different configurations is expected to perform. There is

an emphasis on electricity export, as this will give an indication of how the sugar

mills can expand their additional income stream.

As mentioned previously, the live steam temperature generated for the sugar mill

should not be below 380 °C. Therefore, it was assumed that only solar field outlet

temperatures above 385 °C will be able to generate steam of the necessary quality

for the mill. This resulted in no steam being generated by the solar field just after

sunrise and before sunset, as the solar field could not reach the necessary

temperature during these times of the day. The thermal energy created by the solar

field just after sunrise can, however, be used to start heating the kettle-type heat

exchanger.

Equation 5.2 was used to determine the amount of live steam the solar field can

create. Where 𝑄𝑃𝑇is the heat produced by the solar system, hlsis the enthalpy of

the live steam and hbfwis the enthalpy of the feed water.

�̇�𝑙𝑠 =𝑄𝑃𝑇

ℎ𝑙𝑠−ℎ𝑏𝑓𝑤 (5.2)

If the solar field can only operate during the crushing season, as with

Configuration 1, then 21 820.04 ton live steam can be generated for the normal

sugar mill and 28 817.4 ton for the mill with the refinery, using the east-west

tracking system. For Configuration 1 the solar generated live steam cannot be

used to generate extra electricity, only to ease the boiler’s load, which will save

bagasse or coal. For the generic sugar mill, the ratio of bagasse to live steam is

0.45:1 (Starzak and Davis, 2016). Using this, the amount of bagasse that can

possibly be saved was determined, as shown in Table 15. The amount of coal

saved is determined by using the 1:4 coal to bagasse ratio as specified by Smith et

al. (2016). As in Section 4, this is to better show the added value of the solar

system.

Table 15: Impact of solar live steam generation.

Setup

Bagasse savings Coal savings Extra electricity

generated

Normal

mill [ton]

Mill with

refinery

[ton]

Normal

mill [ton]

Mill with

refinery

[ton]

Normal

mill

[MWhel]

Mill with

refinery

[MWhel]

Configuration 1 9 838.7 12 993.8 2 459.7 3 248.45 - -

Configuration 2 8 964.32 12 290.55 2 241.08 3 072.64 2034.6 2 682.3

Configuration 3 - - - - 5 170.8 6 825.2


73

As mentioned previously, Configuration 2 will allow the solar system to save

bagasse and generate extra electricity. The amount of bagasse which can be saved

is calculated using the same method as for Configuration 1. Table 15 shows that

the amount of bagasse or coal saved is a bit lower than for Configuration 1, this is

because the north-south tracking system is used for Configuration 2.

Configuration 3 can generate the most electricity, because the solar field is used

exclusively for this throughout the whole year, not saving any bagasse.

Table 15 just shows the impact the solar system can make, however, the use of the

different configurations also have an impact on the electricity generation. As

mentioned before, extra electricity can be generated because of the more efficient

CEST, the fact that the let-down steam can now pass through the turbine and the

fact that it can function outside of the crushing season. Equation 5.3 was used to

determine the amount of electricity (W) produced by the turbines and Table 16

shows the values used for these calculations. This was also used to determine the

amount of electricity produced by the solar system as reported in Table 15.

𝑊 = 𝜂𝑂𝑇𝐸 × 𝜂𝑒𝑙 × 𝜂𝑖𝑠 × 𝑚𝑙𝑠(ℎ𝑙𝑠 − ℎ𝑡𝑢𝑟,𝑜𝑢𝑡) (5.3)

Table 16: Values used to determine electricity generation.

Parameter Symbol Value Reference

Overall time efficiency 𝜂𝑂𝑇𝐸 80.83 % Smith et al. 2016

Electrical efficiency 𝜂𝑒𝑙 97 % Giostri et al. 2012

BPST isentropic efficiency 𝜂𝑖𝑠 75 % Foxon 2017

CEST isentropic efficiency 𝜂𝑖𝑠 80 % Ensinas et al. 2007

CEST steam outlet enthalpy ℎ𝑡𝑢𝑟,𝑜𝑢𝑡 2 332.9 kJ/kg* Petchers 2012

* The CEST outlet steam temperature and pressure is taken as 40 °C and 7.34 kPa.

Sugar mills usually do not use all of the electricity it produces, exporting the extra

electricity. If the exported electricity can be sold to the national grid it can become

a major source of additional income as in Brazil (Burin, et al., 2016). Sugar mills

use 22 kWhel per ton of cane crushed (Rein, 2007) and the refinery uses 65 –

70 kWhel per ton of raw sugar refined (Foxon, 2017). The SAM simulations show

that the two solar systems use 275 MWhel and 361 MWhel throughout the year due

to the parasitic loads of the tracking systems and heat transfer pumps. Therefore,

to calculate the amount of electricity that can be exported, the electricity usage of

the mill and solar field is subtracted from the amount of electricity generated as

calculated using Equation 5.3.

Figure 39 shows the possible electricity exports for the various configurations.

Configuration 1 exports the same amount of electricity the mill would have, since

no extra electricity is generated through this setup. Configuration 2 shows a

257 % increase in electricity export for the normal mill and a 102 % increase for


74

the mill with the refinery. Configuration 3 shows a 297 % increase in electricity

export for a normal mill, while the mill with the refinery shows a 111 % increase.

The normal mill sees a higher increase in electricity exports compared to the mill

with the refinery. This is due to the fact that the mill with the refinery was already

exporting a bit of electricity outside of the crushing season, due to the refinery

which still operates then. Therefore, the extra electricity produced and exported

outside of the crushing season due to the CEST and solar field, does not have such

a big impact as with the normal mill.

Figure 39: Possible electricity exports.

0

5000

10000

15000

20000

25000

Generic Config 1 Config 2 Config 3

Ele

ctrc

ity

Exp

ort

ed

[M

Wh

]

Normal Sugar Mill

Sugar Mill withRefinery


75

6 Economic Assessment

This section will discuss various economic factors that need to be considered for

solar thermal integration. It will discuss the levelised cost of heat (LCOH) and

internal rate of return (IRR) for both integration options and how these indicators

can help determined if the integration points are indeed feasible or not. This study

uses Euro for the financial calculations because it is a much more stable currency

than the South African Rand. This will allow for price comparisons in the future,

without worrying too much about the effect of the exchange rate.

6.1 Investment Costs

6.1.1 Specific investment costs of the bagasse drying solar field

In order to obtain reference values for the solar bagasse drying integration point,

the plant database for solar heat for industrial processes (AEE Intec, 2017) was

assessed. By October 2017 the database lists 253 solar systems, unfortunately

only a few of the listings supply any economic information. The Blackdot

database (Blackdot Energy, 2017) unfortunately does not give any economic

information about the listed projects.

The database lists 20 solar systems using air collectors. For this study only

systems larger than 200 m2 were considered, because the suggested solar system

is quite large and smaller plants tend to be more expensive due to the economics

of scale. A larger value was not chosen due to the very limited amount of large

scale systems listed on the database. Of the 10 listed air collectors only two give

any economic information. The specific investment costs differ considerably with

one system costing 73.54 €/m2 and the other 291.67 €/m2, resulting in an average

system cost of 183€/m2, but with major deviation.

This large deviation does not seem uncommon, Wang et al. (2015) mentions that

solar heating costs can vary considerably due factors such as weather conditions,

system complexity and application. Joubert et al. (2016) also mentions that the

solar system costs in South Africa vary considerably due to the factors listed

above and a low level of market maturity.

The two systems mentioned above both use flat plate collectors, so to get a better

idea of the cost of an evacuated tube air collector, the evacuated tube collector

solar systems in the database were also analysed. The database was filtered for

systems larger than 1000 m2. This could be done, because there are much more

registered evacuated tube collector systems than air collector systems in the

database. This should also give a better estimate of the specific investment cost, as

the smaller systems in the database are much more expensive. Of the 7 systems

listed in the database, four of them give economic information. The specific

investment cost vary from 129 €/m2 to 200 €/m2, resulting in an average of


76

168.78 €/m2. It was decided to rather use this value for this study’s financial

calculations as it seems more representative of the simulated solar system and the

fact that the values which have been averaged do not vary as much as with the air

collectors.

6.1.2 Specific investment costs of the live steam generation solar field

Kurup and Turchi (2015) set out a report on parabolic trough collector costs,

specifically focusing on some of the collectors available in SAM. As mentioned in

Section 5.3, FE GmbH Ultimate Trough collectors were used for the simulation,

Kurup and Turchi (2015) calculated that the overall installed cost for the collector

should be 178 $/m2. This is however, for a collector which would use molten salt

as heat transfer fluid, therefore, the collector’s receiver would be more expensive

than for a collector using thermal oil. The cost of a receiver for oil is 3 $/m2 lower

than the receiver for molten salts according to one of the other cost analysis in the

report. Therefore, it can be estimated that the installation cost of the collectors for

this study is 175 $/m2.

The heat transfer fluid system is estimated to cost 70 $/m2, this is the cost for the

oil, piping and hardware costs (Kurup and Turchi, 2015). The site improvements

necessary to install the parabolic troughs are estimated to cost 30 $/m2. Therefore,

the total specific investment cost for the solar field will be 275 $/m2 or

236.5 €/m2. 1

The steam generation system is estimated to cost 4.77 €/kWth (Montes et al.,

2009). By multiplying this number with the thermal energy the simulated field in

SAM can produce and then dividing the answer by the solar field size, it is

possible to represent the steam generation value as 2.65 €/m2. Therefore, the total

specific investment costs for the parabolic trough system is 239.15 €/m2.

6.2 Levelised Cost of Heat

The LCOH is a common measure to asses and compare the costs of renewable

energy projects. It represents the average price per unit of thermal energy

generated throughout the lifetime of the solar system and it is independent of the

value of the energy it replaces. The LCOH compares the project’s annual costs

over its financial life time to the annual yield.

Equation 6.1 (Hess et al., 2016) shows how the LCOH can be calculated. The

annual costs (Cn) include the capital investment costs as well as the operation and

maintenance costs. The annual yield (Qn) refers to the amount of thermal energy

produced by the solar system each year (n).

1The Dollar to Euro exchange rate was taken on 1 November 2017 as $ 1 to € 0.86.


77

𝐿𝐶𝑂𝐻 = ∑

𝐶𝑛

(1+𝑑)𝑛𝑁𝑛=0

∑𝑄𝑛

(1+𝑑)𝑛𝑁𝑛=0

⁄ (6.1)

The project’s financial life span (N) was taken as 20 years, even though the solar

systems can well exceed this. The nominal discount rate (d) was taken as 10 % for

this study as suggested by Beukes et al. (2015) and Hess et al. (2016). This value

is based on the expected weighted average cost of capital of the South African

sugar industry, for 30 % equity and 70 % debt. The equity has an expected return

of 14 % and the debt an interest rate of 8 % with a loan period of 10 years.

For the evacuated tube collector system it was assumed that the annual

maintenance costs will be 1 % of the capital expenditure and the annual

operational cost will be 2 % of the yearly energy yield (Verein Detuscher

Ingenieure, 2004). For the parabolic trough system the annual maintenance and

auxiliary energy costs were assumed at 2 % of the capital costs, as is common for

CSP systems (Hernandez Moro and Martinez-Duart, 2012). These costs are set to

go up by 6 % per annum along with the consumer price index of South Africa, for

both of the systems. The auxiliary electricity tariff imported from the grid is

assumed to be constant at a price of 0.5 ZAR/kWhel (3.05 Euro-ct/kWel)2 (Hess et

al. 2016).

The financial calculations in this study only take the cost of the solar systems into

account, and not the cost of the drying system or the cost of CEST’s for

Configuration 2 and 3. This study assumes that the sugar mills are planning to

implement the suggested systems or configurations in anyway, as it will benefit

the sugar mills. The solar system is only seen as an alternative to using bagasse or

coal.

Hess et al. (2016) calculated the LCOH of coal for the STEP-Bio project. It was

assumed that an existing boiler can burn the necessary amount of South African

coal without having to undergo any alterations or additional costs. Thus, only the

cost of the coal is compared to the cost of the solar system. It was calculated that

coal has a LCOH of 4.03 Euro-ct/kWh, taking into account that the current value

of coal is 1.2 Euro-ct/kWh and will increase with 12.3 % per year. This increase is

based on the coal prices paid by sugar mills after delivery (Hess et al., 2016).

Table 17 shows the calculated LCOH values for the two integration options. The

LCOH of the integration points are significantly lower if they can run throughout

the whole year, as shown in the Annual column. This is simply due to the fact that

by operating the whole year, the system can produce more energy, while the

capital investment stays the same. The LCOH of the mill with the refinery is also

slightly higher than the normal mill. This is because of its marginally higher solar

multiple and subsequent lower system efficiency.

2The Rand to Euro exchange rate was taken on 1 November 2017 as R 1 to € 0.061


78

Table 17: LCOH of the solar thermal integration options.

Integration point Size of solar

field [m2]

Capital

Investment

[Euro]

LCOH [Euro-ct/kWh]

Crushing

season Annual

Solar bagasse drying 22 425 3 784 902 6.77 4.18

Live Steam for normal mill 41 280 9 872 112 8.05 5.17

Live steam for mill with refinery 55 040 13 162 816 8.12 5.22

Despite the annual LCOH values being quite low, they are still higher than the

LCOH of coal, meaning that coal would be a cheaper option to produce the

thermal heat for the processes. However, if the specific investment costs of the

solar systems can be reduced, the LCOH values would be very competitive. As

mentioned in Section 6.1.1, the cheapest air collector system cost 73.54 €/m2, if

this value is used in the calculations, the LCOH for the evacuated tube solar field

would be 1.87 Euro-ct/kWh. Even if the system only functions during the

crushing season, the LCOH would be lower than that of coal, at 3.0 Euro-ct/kWh.

The International Renewable Energy Agency (2016) predicts that the cost of

parabolic trough solar fields will decrease by 23 % by 2025. Therefore the

specific investment cost would be 184 €/m2, resulting in an LCOH of 3.98 Euro-

ct/kWh for the normal mill and 4.02 Euro-ct/kWh for the mill with the refinery if

they were to operate throughout the whole year. Unfortunately, a parabolic trough

system only operating during the crushing season would still not reach a LCOH

close to that of coal.

The proposed carbon tax in South Africa can also have a large influence in the

feasibility of using coal as boiler fuel. A maximum of R 120 per ton of CO2

emitted will have to be payed as carbon tax (Deloitte, 2015); if this is incorporated

into the LCOH calculations of coal, the value would increase to 5.12 Euro-

cent/kWh. Solar thermal integration is much more competitive if this is taken into

account, especially considering the potential future cost reductions of solar

thermal collector technology.

The levelised cost of electricity (LCOE) for the live steam generation integration

point can be calculated in a very similar manner to the LCOH calculations.

Instead of using the annual thermal energy yield, the annual electricity yield is

used. By implementing the annual electricity yield in Equation 6.1, it was

calculated the LCOE for the normal mill would be 25.7 Euro-ct/kWh and

25.9 Euro-ct/kWh for the mill with the refinery. This figure compares relatively

well with the findings of Burin et al. (2016), who calculated that the LCOE of

solar thermal integration into a Brazilian sugar mill would be 22 Euro-ct/kWh. It

is, however, important, to take into account that the annual DNI is higher at the

location considered by Burin et al. (2016) and the central receiver system


79

simulated by the study is considered to cost less than the parabolic trough system

simulated for this study.

6.3 Internal Rate of Return

The IRR estimates the discount rate (d) that would have a zero net present value

(NPV) as a result. Unlike the LCOH, the IRR takes the value of the fuel the solar

system replaces into account. The IRR of each integration point was determined

using Equation 6.2 (Hess et al., 2016), to see if it exceeds the sugar milling

industry’s hurdle rate of 10 - 15 % for projects or investments.

𝑁𝑃𝑉 = ∑𝐶𝑛

(1+𝑑)𝑛𝑁𝑛=0 (6.2)

The value of the coal used by South African sugar mills is taken as R 1 100 per

ton (Hess et al., 2016). In order to calculate the value of the thermal energy

produced by the solar systems, Equation 6.3 was used. The amount of coal saved

and thermal energy delivered to the systems are shown in Table 18 along with the

calculated value of the solar thermal energy and the IRR.

𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑠𝑜𝑙𝑎𝑟 𝑡ℎ𝑒𝑟𝑚𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 = 𝑇𝑜𝑡𝑎𝑙 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑐𝑜𝑎𝑙 𝑑𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑑

𝑇𝑜𝑡𝑎𝑙 𝑡ℎ𝑒𝑟𝑚𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 𝑑𝑒𝑙𝑖𝑣𝑒𝑟𝑒𝑑 (6.3)

Table 18: IRR for the integration options.

Integration

Energy

delivered

[kWhth]

Coal

displaced

[ton]

Value

[Euro-

ct/kWh]

IRR [%]

CS A

Bagasse Drying 6 833 160 1 020.38 1.00 -0.4 4.2

Live steam for normal

mill 15 602 360 2 459.7 1.06 -3.9 2.0

Live steam for mill

with refinery 20 651 280 3 248.45 1.06 -4.0 1.9

From Table 18 it is possible to see that it is imperative that the solar systems run

throughout the whole year, as this makes a considerable difference in the financial

feasibility of the systems. However, the annual figures are still not high enough

for the sugar industry to consider investment, as it falls well below their hurdle

rate.

The solar bagasse drying rate can reach an IRR of 12.9 % if the specific

investment cost is set to the minimum of 73.54 €/m2 and it operates throughout

the whole year. Rotary dryers are quite robust and can be used to dry various

types of materials, this is one of the reasons it was chosen. The drying system can

therefore be used to dry other biomass, like wood chips for example, which is

sometimes used as an auxiliary boiler fuel in sugar mills. The chips can be dried

outside of the crushing season and then stored until it is needed.


80

Unfortunately the solar live steam generation integration systems still cannot

reach the required IRR even when the potential future cost reductions are taken

into account. For this integration option a financial help or subsidies of at least

35 % of the capital cost is needed to reach an IRR above 10 %.

The proposed carbon tax can increase the IRR’s of the integration options,

because the coal it replaces is more expensive. However, it does not make a big

enough difference for any of the integration points to reach the industry’s hurdle

rate. For the solar bagasse drying integration point, the IRR can increase with 2.4

percentage points and the solar live steam integration points can both increase

with 2.6 percentage points.

The electricity generated by the solar energy in the sugar mills can be exported to

the national grid under South Africa’s Renewable Energy Independent Power

Producer Procurement Program (REIPPPP), creating a new income stream for the

sugar mills. To reach an IRR of 15 % the sugar mills would have to receive a

feed-in tariff of 36 Euro-ct/kWh. This is a very high tariff and would most

probably not be accepted by REIPPPP. If the cost reduction measures are taken

into account the feed-in tariff can be lowered to 17.5 Euro-ct/kWh, but this would

still be almost double compared to what South African CSP plants received for

the previous bidding round, Window 3 (Eberhard et al. 2014).

The necessary feed-in tariff is quite high due to two main factors. One is the low

annual DNI of Durban. As mentioned before, the annual sum of DNI at Durban is

1350 kWh/m2, while for normal CSP plants it is significantly higher. Upington,

the centre of South Africa’s CSP developments, receives almost 3000 kWh/m2 of

DNI per year (GeoSUN, 2013), making CSP power plants very feasible there. The

other reason for the high feed-in tariff is the relatively low thermal to electricity

ratio of the sugar mill’s power block, despite considering the CEST. The ratio of

the thermal energy delivered to the sugar mill and the possible electricity it can

generate was calculated to be 20 %. In commercial CSP plants the power block

efficiency can be expected to be above 35 % (Montes et al., 2009).


81

7 Conclusion

The conclusion gives a short summary of the results and highlights the best

integration option in terms of results and economic feasibility. It will also mention

assumptions that were made which influenced the outcome of the study. The

implications of the findings in this study on the sugar milling industry are

discussed and finally recommendations for further work are given.

7.1 Summary of Findings

This study looked at the potential integration of solar thermal energy into South

African sugar mills. The two options researched for this integration is solar live

steam generation and solar bagasse drying. Both of the integration points were

developed further from previous STEP-Bio studies done by Beukes et al. (2015)

and Hess et al. (2016). Detailed simulations of the solar fields were created in

order to determine the impact they could make on a generic sugar mill. The

simulation results were also used to complete an economic assessment to

determine the financial feasibility of the two integration points.

The simulations show that the solar systems can make a relatively sizeable impact

despite the sugar mill’s energy consumption significantly surpassing their

capabilities. However, the simulation results did point out a few concerns. For the

bagasse drying integration point, the relatively low efficiency of the air collectors

hampered the solar system’s performance and for the live steam generation point

the low annual DNI of Durban greatly limits the performance of the solar system.

The bagasse drying system would allow the mill to reduce its bagasse usage by

5 %. By integrating the solar thermal energy system the decrease in bagasse usage

can be lowered to 7.05 %, making a significant contribution compared to when

only exhaust steam is used. The solar live steam generation integration point will

allow for bagasse/coal saving and extra electricity generation. Configuration 1 can

save the most coal, 2 459.7 ton for a normal mill and 3 248.4 ton for a mill with a

refinery. Configuration 3 can generate the most electricity and would enable to

increase the electricity exported by 297 % for a normal mill and by 111 % for a

mill with a refinery. It is, however, Configuration 2 that can make the largest

impact, as it can save almost as much coal and produce as much electricity as

Configuration 1 and 3, respectively.

The economic assessment shows that under current conditions the bagasse drying

integration point offers the lowest LCOH and highest IRR. However, even for this

integration point, the solar system would still be more expensive than using coal

as fuel to produce thermal energy. The IRR is also lower than the hurdle rate set

by the sugar milling industry. The economic assessment underlines the

importance of the solar systems operating throughout the whole year, showing


82

that there would be no return on investment if the systems were only to operate

during the crushing season.

There is some uncertainty regarding the economic calculations due to varying

reported investment costs, potential reductions in parabolic trough collector costs

and the increase of coal costs due to planned carbon taxation in South Africa. If

the carbon tax is taken into account the solar systems would be very competitive,

as the LCOH of coal would be quite close to that of the solar systems’. If the

lowest prices for solar integration are taken into account the LCOH of both

integration options will be significantly lower than that of coal. However, despite

the low LCOH values that can be reached, the IRR values still stay relatively low

compared to the hurdle rate set by the sugar industry. Only the bagasse drying

integration point will be able to achieve an IRR higher than 10 % if the lower

investment costs are taken into account.

If the sugar mills can export the extra electricity that can be generated to the

national grid, a feed-in tariff of 36 Euro-ct/kWh is needed to achieve an IRR of

15 %. This is an extremely high price if it is compared to what current CSP power

projects receive as feed-in tariff. If the reduced collector prices are considered the

tariff can be lowered to 17.5 Euro-ct/kWh, which is close to what CSP plants

received in the REIPPPP bidding Window 1, but almost double what they

received in bidding Window 3. The high needed feed-in tariffs can be attributed to

the low annual DNI of Durban and the low thermal to electricity efficiency of the

sugar mills compared to commercial power plants.

7.2 Concluding Remarks

Solar thermal integration into South African sugar mills is technically possible,

allowing the mills to save bagasse/coal and generate extra electricity. It is,

however, not yet financially feasible, with coal being able to provide thermal

energy at a lower cost.

The economic assessment shows that the solar bagasse drying integration point

would be the best option to pursue due to its lower LCOH and higher IRR. It

would also require less capital investment due to a smaller solar field being

needed. The evacuated tube collectors are also better suited for Durban’s solar

resource as it uses GTI, which is higher than the DNI in Durban. The collectors

also require less maintenance and are considerably less complicated to operate

compared to parabolic trough collectors.

The success of the solar live steam generation point will depend on the feed-in

tariff that can be negotiated between the sugar milling industry and Eskom. The

chances of the needed feed-in tariff being paid seems to be rather slim due to how

high it is in comparison with the latest CSP power plants and Eskom’s reluctance

to sign IPP agreements.


83

The future of the solar thermal integration, however, looks bright as collector

costs are set to decrease and the price of using coal increases due to carbon

taxation. If both of these factors come into play solar integration can become very

feasible and should always be considered as a potential thermal energy source.

7.3 Recommendations for Further Work

This study was based on the generic sugar mill described in the BRTEM model.

In order to fully realise the potential impact solar thermal integration can make, an

actual sugar mill will have to be analysed. Some of the existing sugar mills are

located in areas with a higher solar resource and may be better suited for solar

thermal integration due to alternative boiler house configurations which can allow

for a higher boiler turn down ratio.

Some mills have lower steam parameters, which opens the door for thermal

storage systems. A new silicon based heat transfer fluid, developed by

HELISOL®, has recently been introduced and can operate at temperatures up to

425 °C (Zoschke et al., 2017). This will also allow the integration system to

incorporate solar thermal storage. Therefore, the impact a small – medium sized

thermal storage system can make should also be researched. It would help the

solar system deal with transient conditions and allow the solar system to provide

thermal energy after dark. This can make up for the days with very low to no DNI

when the solar system cannot produce any useful thermal energy.

To reach lower LCOH’s and IRR’s for the solar live steam generation integration

point alternative collector technologies must researched. Linear Fresnel systems

and direct steam generation offer the lowest investment cost and according to

Peterseim (2014) is the best option for hybridisation and cogeneration. However,

the uncertainty of this technology’s performance will also have to be taken into

account and how potential financiers would react to it.

In order to get a more accurate value of the thermal requirements of the bagasse

drying system further research will have to be done on the drying kinetics of

bagasse in a rotary dryer. A more detailed drying simulation will have to be done

which takes all the aspects of a rotary dryer into consideration, like the speed of

rotation, number of flights and tilt of the dryer. This will give a clearer picture of

how effective a rotary dryer will be in drying bagasse and if exhaust steam and

solar thermal energy can indeed deliver the needed thermal energy. Furthermore,

the precise effect of the dry bagasse on the boiler efficiency will have to be

calculated using boiler design calculations. This will allow for more accurate

estimations on the effect of bagasse drying on the sugar mill.


84

Appendix A: Airwasol Brochure


85


86


87

Appendix B: Matlab Model Calculations

The following polynomial equations from Kröger 1998) were used to determine

certain thermo-physical properties of air, which was needed to calculate the heat

transfer characteristics in the evacuated tube. Note that all the temperatures should

be entered in Kelvin.

Kinematic viscosity:

𝜈 =(2.287973 × 10−6) + (6.259793𝑇𝑓 × 10−8) − (3.131956𝑇𝑓

2 × 10−11) + (8.15038𝑇𝑓3 × 10−15)

𝑃/(287.08𝑇𝑓)

(A.1)

Where Tf is the film temperature around the glass tube and is seen as the average

between the ambient air temperature and the glass temperature. For the thermal

conductivity of the film the following equation was used:

𝑘 = (−4.937787 × 10−4) + (1.018087𝑇𝑓 × 10−4) − (4.627937𝑇𝑓

2 × 10−8) + (1.250603𝑇𝑓3 × 10−11)

(A.2)

To determine the thermal conductivity and kinematic viscosity of the air inside the

absorber tube, Tf must be replaced with bulk temperature of the air (Tb) inside the

absorber tube. The bulk temperature as also used to determine the specific heat

capacity of the air inside the absorber tube.

𝑐𝑝 = (1.9327𝑇𝑏

4 × 10−10) − (7.9999𝑇𝑏3 × 10−7) + (1.1407𝑇𝑏

2 × 10−3) − (4.4890𝑇𝑏 × 10−1)

+ (1.0575 × 103) (A.3)

Equation 4.9 was derived from Cengel&Ghajar (2015) which gives a formula for

the radiation heat transfer between two concentric cylinders. The formula from

Cengel and Ghajar (2015) is:

𝑄𝑎𝑔 = 𝐴𝑎 𝜎 (𝑇𝑎

4−𝑇𝑔4)

1

𝜀𝑎+

1−𝜀𝑔

𝜀𝑔(

𝑟𝑎𝑟𝑔

) (A.4)

For the Matlab model, the above equation needs to be written in the following

form:

𝑄𝑎𝑔 = (𝑇𝑎−𝑇𝑔)

𝑅𝑎𝑔 (A.5)

Therefore, the thermal radiation resistance (Rag) can be written as:

𝑅𝑎𝑔 =

1

𝜀𝑎+

1−𝜀𝑔

𝜀𝑔(

𝑟𝑎𝑟𝑔

)

𝜎 𝐴𝑎(𝑇𝑎2+𝑇𝑔

2)(𝑇𝑎+𝑇𝑔) (A.6)


88

To calculate the glass temperature (Equation 4.10) the following energy balance

was set up:

𝑄𝑟𝑎𝑑,𝑎 + 𝑄𝑎𝑔 − 𝑄𝑐𝑜𝑛𝑣 − 𝑄𝑟𝑎𝑑𝑠𝑘𝑦 = 0 (A.7)

This can be rewritten as:

𝑄𝑟𝑎𝑑,𝑎 +𝑇𝑎−𝑇𝑔

𝑅𝑎𝑔−

𝑇𝑔−𝑇∞

𝑅𝑔∞−

𝑇𝑔−𝑇𝑠𝑘𝑦

𝑅𝑔𝑠𝑘𝑦= 0 (A.8)

Now all that needs to be done to find Equation 4.10 is to isolate the Tg on the right

hand side of the equation:

𝑇𝑔 =𝑄𝑟𝑎𝑑,𝑎+

𝑇𝑎𝑅𝑎𝑔

+𝑇∞

𝑅𝑔∞+

𝑇𝑠𝑘𝑦

𝑅𝑔𝑠𝑘𝑦1

𝑅𝑎𝑔+

1

𝑅𝑔∞+

1

𝑅𝑔𝑠𝑘𝑦

(A.9)

In the equations above Qrad,a represents the radiation from the sun which the glass

absorbs. This can be determined using Equation A.10:

𝑄𝑟𝑎𝑑,𝑎 = 𝐼𝑠(1 − 𝜏 − 𝜌𝑔) (𝐴𝑔

2) (A.10)

Here τ and ρg represent the transmissivity and reflectivity of the glass. The

reflectivity was taken as 5 %, while looking at Figure 40, it was decided to be

conservative and assume that the glass only lets wave lengths between 0.3 –

2.8 μm through and for this waveband the transmissivity was taken as 92 %.

Borofloat 33 is a Borosilicate glass, very similar to what was used for the

Airwasol collector.

Figure 40: The transmissivity of Borofloat 33 over various wavelengths (Schott, 2017).


89

The following calculations show how much of the solar energy the glass actually

transmits, it was determined using values from Schott (2017) and as set out by

Kalogirou (2009). Firstly, the energy that is emitted from the sun for each of the

specified wavelengths is calculated and then presented as a fraction of the total

emissive power from the sun. Kalogirou (2009), conveniently set up a table where

this fraction is specified as a function of the blackbody’s temperature (in this case

the sun at 5760 K) and the specified wavelengths.

𝜆1𝑇 = 0.3 × 5760 = 1728 𝜇𝑚𝐾

𝜆2𝑇 = 2.8 × 5760 = 16128 𝜇𝑚𝐾

The following values were read from the table specified above:

𝐸𝑏(0 → 𝜆1𝑇)

𝜎 𝑇4= 3.166 %

𝐸𝑏(0 → 𝜆2𝑇)

𝜎 𝑇4= 97.43 %

Therefore, the fraction of solar energy that can be transmitted in the specified

wavelengths is:

𝐸𝑏(𝜆1 → 𝜆2𝑇)

𝜎 𝑇4= 97.43 − 3.166 = 94.264 %

Now the percent of solar radiation transmitted through the glass can be

determined:

𝜏 = 0.92 × 0.94264 = 86.72 %

The irradiance that finally passes through the glass is then used to heat up the

absorber. The following energy balance was set up to eventually derive a method

of calculating the absorber temperature (Equation 4.13):

𝑄𝑟𝑎𝑑 − 𝑄𝑎𝑔 − 𝑄𝑐𝑜𝑛𝑣,𝑎𝑖𝑟 = 0 (A.11)

This can then be rewritten as:

𝑄𝑟𝑎𝑑 −𝑇𝑎−𝑇𝑔

𝑅𝑎𝑔−

𝑇𝑎−𝑇𝑎𝑖𝑟

𝑅𝑎𝑖𝑟= 0 (A.12)

Tanow needs to be isolated on the right hand side of the equation in order to end

up with the same equation as Equation 4.13.


90

𝑇𝑎 = 𝑄𝑟𝑎𝑑+

𝑇𝑔

𝑅𝑎𝑔+

𝑇𝑎𝑖𝑟𝑅𝑎𝑖𝑟

1

𝑅𝑎𝑔+

1

𝑅𝑎𝑖𝑟

(A.13)

In Equation A.12, Qrad represents the irradiance that falls on the absorber. This

can be determined by taking the absorber area exposed to the sun into account

along with its absorptance and the irradiance that comes through the glass tube, as

can be seen Equation A.14.

𝑄𝑟𝑎𝑑 = 𝛼𝐼𝑠 (𝐴𝑎

2) (A.14)

The temperature of the air leaving the 1 cm absorber segment under consideration

can now finally be determined. Equation 14 was derived from the following

energy balance:

�̇� 𝑐𝑝(𝑇𝑎𝑖𝑟,𝑖+1 − 𝑇𝑎𝑖𝑟,𝑖) =𝑇𝑎−𝑇𝑎𝑖𝑟

𝑅𝑎𝑖𝑟 (A.15)

This can then be simplified to Equation 14:

𝑇𝑎𝑖𝑟,𝑖+1 = 𝑇𝑎𝑖𝑟,𝑖 + 𝑇𝑎−𝑇𝑎𝑖𝑟

𝑅𝑎𝑖𝑟 �̇� 𝑐𝑝,𝑎𝑖𝑟 (A.16)


91

280

300

320

340

360

380

400

420

440

0 10 20 30 40 50 60

Ou

tlet

Tem

per

atu

re [

K]

Flow Rate [m^3/h]

Outlet Temperature

Inlet Air Temperature

Appendix C: Matlab Model Comparison

Unfortunately Paradis et al. (2015) et al does not give any results regarding the

incremental temperature rise over the tube length during their experiment. They

focus on the outlet temperatures of the collector tube and if they can model it

correctly as the irradiation, ambient temperature and flow rate changes. They

validate their model using the experimental results shown in the thesis.

Their model shows that the flow rate through the collector tube has the largest

influence on the outlet air temperature. To show the similarity between their

model and the model used for this study, the influence of the flow rate on the

outlet air temperature and the collector efficiency is shown for both models in

Figure 41 - 44:

Figure 41: Change in collector outlet temperature as flow rate changes as predicted by model

of Paradis et al. (2015)

Figure 42: Change in collector outlet temperature as flow rate changes as predicted by model

used in thesis.


92

45

50

55

60

65

70

75

0 10 20 30 40 50 60

Effi

cien

cy [

%]

Flow Rate [m^3/h]

Figure 43: Change in collector efficiency as flow rate changes as predicted by model of

Paradis et al. (2015)

Figure 44: Change in collector efficiency as flow rate changes as predicted by model used in

thesis.

The reason the results for the model used in this study does not go below a flow

rate of 5 m3/h, is because below this flow rate the Reynold’s number is not high

enough so that Gnielinski’s equation can be used accurately. The comparison of

the graphs show that the models’ results are very similar and that the heat transfer

modelling was, therfore, done correctly.


93

Appendix D: Aircow Optimisation

Aircow has an optimisation function which determines the optimum air flow

through the collector rows in the solar field. The program determines this by

considering the system set out in Figure 45.

Figure 45: Aircow’s Global system considered for mass flow optimisation as shown in the

Aircow manual (Fraunhofer ISE, 2017).

The program has four different mass flow optimisation options. The first option is

the optimisation of the mass flow for net primary energy savings. The second

option is to optimise the mass flow to maximise net monetary savings, which is

determined using the value of the auxiliary electricity and the heating medium

replaced. The third option is to optimise the flow for maximum net site power

savings, while the fourth option maximises the net system power savings.

Equations B.1 – B.4 show how the power savings and cost savings were

determined for the various options.

Option 1: 𝑃𝑠𝑎𝑣𝑒𝑑 𝑝𝑒𝑟 𝑓𝑖𝑒𝑙𝑑 𝑎𝑟𝑒𝑎

= �̇�𝑓𝑖𝑒𝑙𝑑 (𝑓𝑝𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑑 𝑒𝑛𝑒𝑟𝑔𝑦 𝑐𝑎𝑟𝑟𝑖𝑒𝑟

𝐴𝑓𝑖𝑒𝑙𝑑×𝜂𝑠𝑦𝑠𝑡𝑒𝑚 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑑) − 𝑃𝑓𝑎𝑛 (

𝑓𝑝𝑎𝑢𝑥𝑖𝑙𝑖𝑎𝑟𝑦 𝑒𝑙𝑒𝑐𝑡𝑟𝑐𝑖𝑡𝑦

𝐴𝑓𝑖𝑒𝑙𝑑)

(B.1)


94

Option 2: 𝐶𝑜𝑠𝑡𝑠𝑠𝑎𝑣𝑒𝑑 𝑝𝑒𝑟 𝑓𝑖𝑒𝑙𝑑 𝑎𝑟𝑒𝑎

= �̇�𝑓𝑖𝑒𝑙𝑑 (

𝑐𝑜𝑠𝑡𝑠 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑑 𝑒𝑛𝑒𝑟𝑔𝑦 𝑐𝑎𝑟𝑟𝑖𝑒𝑟


𝑐𝑜𝑠𝑡 𝑎𝑢𝑥𝑖𝑙𝑖𝑎𝑟𝑦 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦


(B.2)

Option 3: 𝑃𝑠𝑎𝑣𝑒𝑑 𝑠𝑖𝑡𝑒 𝑝𝑜𝑤𝑒𝑟 𝑝𝑒𝑟𝑓𝑖𝑒𝑙𝑑 𝑎𝑟𝑒𝑎

= �̇�𝑓𝑖𝑒𝑙𝑑 (1


1


(B.3)

Option 4: 𝑃𝑠𝑎𝑣𝑒𝑑 𝑠𝑦𝑠𝑡𝑒𝑚𝑝𝑜𝑤𝑒𝑟 𝑝𝑒𝑟𝑓𝑖𝑒𝑙𝑑 𝑎𝑟𝑒𝑎

= �̇�𝑓𝑖𝑒𝑙𝑑 (1

𝐴𝑓𝑖𝑒𝑙𝑑) − 𝑃𝑓𝑎𝑛 (

1


(B.4)

After the amount of power or cost saved is calculated, Aircow uses an algorithm

which considers three different mass flows to determine the maximum of a

quadratic curve fitted to the power or cost function. The range of mass flows will

be reduced step by step and it will start to focus on the predicted maximum until a

high enough precision is reached.

Now that the optimal mass flow rate is known, the amount of collectors in a row

can be determined. Aircow can use the efficiency map it created to determine the

temperature increase per collector. This is then used along with the user defined

solar field exit temperature to determine how much collectors in a row is

necessary to gain the necessary temperature lift. The answer Aircow gives almost

always has a decimal in it, the problem is that in reality there is no such thing as a

fraction of a collector, only an integer amount of collectors can be used.

Therefore, the decimal needs to be rounded up to the nearest integer and the

amount of rows in the system needs to be adjusted so that the thermal power at the

field outlet is more or less the same or the necessary mass flow rate through the

total system is reached. Figure 46 gives a visual representation of this process.

The system calculations now need to be repeated with the newly specified

collector rows and optimised mass flow rate in order to determine how the “real”

system would perform. The amount of collectors calculated initially can also be

rounded down, the system performance of the two different collector rows should

then be compared to one another as in Figure 46.

To determine the price of the energy replaced the following calculations were

made, taking into account that the replaced heating medium is exhaust steam:

1 𝑘𝑊ℎ = 3600 𝑘𝐽

To calculate the amount of steam, the power is divided by the enthalpy difference

of the exhaust steam and the condensate that forms after it heats the air.


95

𝑚 =3600

2707.95 − 504.68= 1.63 𝑘𝑔

The ratio between exhaust steam and live steam is 1.2:1. While the ratio between

live steam and bagasse is 1:0.45, then lastly the ratio for bagasse to coal is 1:0.25

(Starzak and Zizhou, 2015).

1.63 𝑘𝑔 𝑒𝑥ℎ𝑎𝑢𝑠𝑡 𝑠𝑡𝑒𝑎𝑚 = 1.36 𝑘𝑔 𝑙𝑖𝑣𝑒 𝑠𝑡𝑒𝑎𝑚

1.36 𝑘𝑔 𝑙𝑖𝑣𝑒 𝑠𝑡𝑒𝑎𝑚 = 0.614 𝑘𝑔 𝑏𝑎𝑔𝑎𝑠𝑠𝑒

0.614 𝑘𝑔 𝑏𝑎𝑔𝑎𝑠𝑠𝑒 = 0.153 𝑘𝑔 𝑐𝑜𝑎𝑙

One ton of coal costs R 1 100 (Hess et al., 2016), therefore:

0.153 𝑘𝑔 𝑐𝑜𝑎𝑙 = 𝑅 0.1688 𝑜𝑟 16.88 𝑐

Therefore, for each kWh of heat replaced by the solar system, 16.88 c of coal is

saved.


96

Figure 46: Comparison of different number of collectors in a row with optimised flow as

shown in the Aircow manual (Fraunhofer ISE, 2017).


97

0

5

10

15

20

25

Sola

r Th

erm

al O

utp

ut

[MW

]

N - STrackingSystem

E - WTrackingSystem

0

5

10

15

20

25

Sola

r Th

erm

al O

utp

ut

[MW

]

N - STrackingSytem

E - WTrackingSystem

Appendix E: Comparison of Tracking Systems

This appendix shows the difference in the daily outputs of the two different

tracking systems. As mentioned in Section 5.4 the output of the east-west tracking

system during the day is uneven, while the output of the north-south tracking

system varies throughout the year. Figures 47 show the performance of the two

systems on the day of the winter solstice. This shows that the north-south tracking

system performs rather poor and why it was considered using the east-west

tracking system for the crushing season which stretches throughout the winter.

Figure 47: Comparison of the performance of the two tracking systems on 21 June.

Figure 48 shows the two systems on the day of the summer solstice, when the

north-south system performs considerably better. The east-west system’s

performance is much the same as it was on the winter solstice, this is one of the

system’s advantages, there is little difference in its performance during different

seasons.

Figure 48: Comparison of the performance of the two tracking systems on 20 December.


98

References

Abbas, R., Montes, M. J., Rovira, A., & Martı´nez-Val, J. M. (2016). Parabolic

trough collector or linear Fresnel collector? A comparison of optical

features including thermal quality based on commercial solutions. Solar

Energy, 198-215.

AEE Intec. (2017, 10 14). Plants Database. Retrieved from SHIP: http://ship-

plants.info/

Amos, W. A. (1998). Report on Biomass Drying Technology. National Renewable

Energy Laboratory.

Backen, M., Ferruzza, D., Larsen, L. K., & Haglind, F. (2017). Hybridization of

Concentrated Solar Power and Biomass Combustion for Combined Heat

and Power Generation in Northern Europe. SolarPACES. Santiago .

Baharoon, D. A., Rahman, H. A., Omar, W. Z., & Fadhl, S. O. (2015). Historical

development of concentrating solar power technologies to generate clean

electricity efficiently – A review. Renewable and Sustainable Energy

Reviews, 996-1027.

Bajura, R. A., & Jones, E. H. (1976). Flow Distribution Manifolds. Journal of

Fluids Engineering , 654-666.

Bakshi, A. S., & Singh, R. P. (1980). Drying characteristics of parboiled rice.

International Symposium on Drying .

Beukes, H. T., Brent, A. C., & Hess, S. (2015). Opportunities for Solar Thermal

Process Heat Integration in South African Sugar Mills. Stellenbosch

University Faculty of Engineering .

Bhatt, S. (2014). Concentrating solar power in cane-based sugar cogen plants: a

strategic energy option for India. Electrical and Power Review, 30-33.

Biomass Knowledge Portal. (2015). First Commercial CSP-Biomass Hybrid Plant

in Spain. New Delhi: Ministry of New and Renewable Energy, India.

Blackdot Energy. (2017, 10 15). SA's Most Amazing Solar Thermal Projects.

Retrieved from Blackdot Energy: http://www.blackdotenergy.co.za

Bruce, D. M., & Sinclair, M. S. (1996). Thermal Drying of Wet Fuels:

Opportunities and Technology. Vancouver : H.A. SIMONS LTD.

Burin, E. K., Vogel, T., Multhaupt, S., Thelen, A., Oeljeklaus, G., Gorner, K., &

Bazzo, E. (2016). Thermodynamic and economic evaluation of a solar

aided sugarcane bagasse cogeneration power plant. Energy , 416-428.

Cabrera , F. J., Fernandez-Garcia , A., Silva, R. M., & Perez-Garcia, M. (2013).

Use of parabolic trough solar collectors for solar refrigeration and air-


99

conditioning applications. Renewable and Sustainable Energy Reviews,

103-118.

Calamateo, D., & Zhou, D. (2015). A Pre-Feasibility Study of a Small-Scale

Concentrating Solar Power Plant for an Industrial Application in

Sacramento, California. Proceedings of the 2nd International Conference

on Fluid Flow, Heat and Mass Transfer, (pp. 180-1 - 180-8). Ottawa,

Canada.

Cardemil, J. M., Starke, A. R., Scariot, V. K., Grams, I. L., & Colle, S. (2013).

Evaluating solar radiation attenuation models to assess the effects of

climate and geographical location on the heliostat field efficiency in

Brazil. SolarPACES, (pp. 1288-1297). Las Vegas .

Çengel, Y. A., & Cimbala, J. M. (2010). Fluid Mechanics: Fundamentals and

Applications. New York: McGraw-Hill.

Çengel, Y. A., & Ghajar, A. J. (2015). Heat and Mass Transfer: Fundamentals

and Applications - Fith Edition. New York: McGraw-Hill Education.

Cook, E. M. (1991). Process Drying Practice. New York : McGraw-Hill.

CSP World Organisation. (2015). CSP World Map. Retrieved October 16, 2017,

from CSP World: http://cspworld.org/cspworldmap/kimberlina-stpp

De Oliveira, L. F., Correa, J. L., Tosato, P. G., Borges, S. V., Alves, J. G., &

Fonseca, B. E. (2011). Sugarcane Bagasse Drying in a Cyclone: Influence

of Device Geometry and Operational Parameters. Drying Technology, 29 ,

946–952.

Deloitte. (2015). Draft Carbon Tax Bill Released. Retrieved November 3, 2017,

from Deloitte :

https://www2.deloitte.com/za/en/footerlinks/pressreleasespage/carbon-tax-

bill.html

Don, C. E., Mellet, P., Ravno, B. D., & Bodger, R. (1977). Calorific Values of

South African Bagasse. South African Sugar Technologists' Association,

(p. 5). Durban.

Duffie, J. A., & Beckman, W. A. (2006). Solar Engineering of Thermal Processes

. New Jersey : John Wiley & Sons .

Eberhard, A., Kolker, J., & Leigland, J. (2014). South Africa’s Renewable Energy

IPP Procurement Program: Success Factors and Lessons. Washington,

DC: Public-Private Infrastructure Advisory Facility.

Electromagnetic Foundations of Solar Radiation Collection. (2017, July 10). Solar

Energy. Retrieved October 16, 2017, from Energy Professional

Symposium: http://energyprofessionalsymposium.com/?p=19152


100

Ensinas, A. V., Nebra, S. A., Lozano, M. A., & Serra, L. M. (2007). Analysis of

process steam demand reduction and electricity generation in sugar and

ethanol production from sugarcane. Energy Conversion & Management,

2978-2987.

Fernandez-Garcia, A., Zarza, E., Valenzuela, L., & PerezM. (2010). Parabolic-

trough solar collectors and their applications. Renewable and Sustainable

Energy Reviews, 1695-1721.

Foxon, K. (15 May 2017). Personal communication between Willem Krog

(Stellenbosch University) and Katherine Foxon (SMRI) via email.

Foxon, K. (22 November 2017). Personal communication between Willem Krog

(Stellenbosch University) and Dr Katherine Foxon (SMRI) via email.

Fraunhofer ISE. (2017). Aircow Version 3.00 Manual . Freiburg: Fraunhofer ISE.

Fredrikson, R. W. (1984). Utilisation of Wood Waste as Fuel for Rotary and Flash

Tube Wood Dryer Operation. Biomass Fuel Drying Conference, (pp. 1-

16).

Freire, F., Figueiredo, A., & Ferrao, P. (2001). Postharvest Technology:

Modelling High Temperature, Thin Layer, Drying Kinetics of Olive

Bagasse. Journal of Agricultural Engineering Research, 397-406.

GeoSUN. (2012). KZN Solar Maps & GIS Layers. Retrieved October 17, 2017,

from GeoSUN: http://geosun.co.za/downloads/kzn-solar-maps-gis-layers-

2012/

GeoSUN. (2013). South African GIS Layers. Retrieved November 3, 2017, from

GeoSUN: http://geosun.co.za/downloads/south-african-gis-layers/

Giglio, A., Lanzini, A., Leone, P., García, M. R., & Moya, E. Z. (2017). Direct

steam generation in parabolic-trough collectors: A review about

thetechnology and a thermo-economic analysis of a hybrid system.

Renewable and Sustainable Energy Reviews, 453-473.

Giostri, A., Binotti, M., Astolf, M., Silva, P., Macchi, E., & Manzolini, G. (2012).

Comparison of different solar plants based on parabolic trough technology.

Solar Energy, 1208-1221.

Gong, Z.-X., & Mujumdar, A. S. (2008). Software for Design and Analysis of

Drying Systems . Drying Technology, 884-894.

Gunther, M., Joemann, M., & Csambor, S. (2013). Parabolic Trough Technology.

In D. Institute of Solar Research, Advanced CSP Teaching Materials (p.

Chapter 5). enerMENA.


101

Hanrieder, N., Wilbert, S., Mancera-Guevara, D., Buck, R., Giuliano, S., & Pitz-

Paal, R. (2017). Atmospheric extinction in solar tower plants – A review.

Solar Energy, 193-207.

Heller, L. (2013). Literature Review on Heat Transfer Fluids and Thermal Energy

Storage Systems in CSP Plants. Stellenbosch : Soler Thermal Research

Group, Stellenbosch University .

Hernandez Moro, J., & Martinez-Duart, J. M. (2012). SP electricity cost evolution

and grid parities based on the IEA roadmaps. Energy Policy, 184-192.

Hess, S. (11 July 2016). Personal Communication between Willem Krog and Dr

Stefan Hess at Stellenbosch University .

Hess, S., Beukes, H., Dinter, F., & Smith, G. (2016). Initial Study on Solar

Process Heat for Sout African Sugar Mills. South African Sugar

Technologists' Association Congress, (pp. 324-349). Durban.

Ho, C. K. (2017). Advances in central receivers for concentrating solar

applications. Solar Energy, 38-56.

Hugot, E. (1972). Handbook of cane Sugar Engineering . Elsevier Publishing

Company .

Iftekhar Hussain, C. M., Norton, B., & Duffy, A. (2017). Technological

assessment of different solar-biomass systems for hybrid power generation

in Europe. Renewable and Sustainable Energy Reviews, 1115-1129.

Ilchmann, C., Klatt, M., & Hess, S. (2016). Modeling of Solar Process Heat

Systems in an Opern Source Simulation Environment. Cottbus:

Brandenburgische Technische Universität.

IRENA. (2012). Renewable Energy Technologies: Cost Analysis Series.

International Renewable Energy Agency.

IRENA. (2016). The Power to Change: Solar and Wind Cost Reduction Potential

to 2025. International Renewable Energy Agency.

Joubert, E. C., Hess, S., & Van Niekerk, J. L. (2016). Large-scale solar water

heating in South Africa: Status, barriers and recommendations .

Renewable Energy, 809-822.

Kalogirou, S. A. (2009). Solar Energy Engineering: Processes and Systems .

Burlington : Elsevier .

Keey, R. B. (1994). Progress in Understanding Drying. In R. B. Keey, & V.

Rudolph, Drying '94' (pp. 75-88). Univeristy of Queensland .

Kemp, I. C., & David, E. O. (2002). Modelling of Particulate Drying in Theory

and Practice . Drying Technology, 1699-1750.


102

Kotze, C. (2016). Four Ways of Improving Boiler Efficiency . South African

Sugar Technologists Association Congress, (pp. 498-507). Durban.

Kraemer, S. (2017, August 25). SolarReserve Breaks CSP Price Record with 6

Cent Contract. Retrieved 2 September, 2017, from SolarPACES:

http://www.solarpaces.org/solarreserve-breaks-csp-price-record-6-cent-

contract/

Kröger, D. G. (1998). Air-Cooled Heat Exchangers and Cooling Towers .

Stellenbosch: University of Stellenbosch.

Kurup, P., & Turchi, C. S. (2015). Parabolic Trough Collector Cost Update for

the System Advisor Model (SAM). Alexandria, USA: National Renewable

Energy Laboratory.

Laubscher, R. (2017). Personal communication between Ryno Laubscher and

Willem Krog at Stellenbosch University.

Liu, M., Tay, N. H., Bell, S., Belusko, M., Jacob, R., Will, G., . . . Bruno, F.

(2016). Review on concentrating solar power plants and new

developments in high temperature thermal energy storage technologies.

Renewable and Sustainable Energy Reviews, 1411-1432.

Madho, S., Davis, S. B., & Bhyrodeyal, L. (2017). Ninety-Second Annual review

of the Milling Season in Shouthern Africa (2016/2017). South African

Sugar Technologists’ Association Congress, (pp. 20-50). Durban.

Magasiner, N. (1987). The Effect of Fuel Moisture Content on the Performance of

a Typcial Bagasse Fired Watertube Boiler. The South African Sugar

Technologists' Association, (pp. 86-89).

Magasiner, N., Van Alphen, C., Inkson, M., & Misplon, B. (2001). Characterising

Fuels For Biomass – Coal Fired Cogeneration . South African Sugar

Technologists' Association, (p. 17). Durban .

Masters, K. (1985). Spray Drying Handbook. New York: John Wiley & Sons.

Meyer, R. (2016). Solar Resource Assessment/Mapping. Stellenbosch :

STERG/GeoSun.

Miles, T. R., Miles, T. R., Baxter, L. L., Bryers, R. W., Jenkins, B. M., & Oden,

L. L. (1995). Alkali Deposits Found in Biomass Power Plants. National

Renewable Energy Laboratory .

Montes, M. J., Abandes, A., Martinez-Val, J. M., & Valdes, M. (2009). Solar

multiple optimization for a solar-only thermal power plant, using oil as

heat transfer fluid in the parabolic trough collectors. Solar Energy, 2165-

2176.


103

Mujumdar, A. S. (1995). Handbook of Industrial Drying Second Edition. New

York: Marcel Dekker.

Mujumdar, A. S., Borde, I., & Levy, A. (2006). Handbook of Industrial Drying,

Third Edition. CRC Press.

Munoz-Anton, J., Biencinto, M., Zarza, E., & Diez, L. E. (2014). Theoretical

basis and experimental facility for parabolic trough collectors at high

temperature using gas as heat transfer fluid. Applied Energy, 373-381.

National Renewable Energy Laboratory . (2017, January 17). System Advisor

Model Version 2017.1.17. Retrieved April 7, 2017, from National

Renewable Energy Laboratory : https://sam.nrel.gov/download

National Renewable Energy Laboratory. (2013). Concentrating Solar Power

Projects: Borges Termosolar. National Renewable Energy Laboratory of

the U.S. Department of Energy.

National Renewable Energy Laboratory. (n.d.). Concentrating Solar Power

Projects. Retrieved October 19, 2017, from National Renewable Energy

Laboratory: https://www.nrel.gov/csp/solarpaces/

Naude, D. P., McIntyre, P. J., & Field, S. J. (1993). The Design and Operation of

Boiler Plant Utilising Furfural Residue as a Fuel . South African Sugar

Technologists' Association, (p. 5). Durban.

Nixon, J. D., Dey, P. K., & Davies, P. A. (2012). The feasibility of hybrid solar-

biomass power plants in India. Energy, 541-554.

Oliverio, J. L., Avila, A. C., Faber, A. N., & Soares, P. A. (2014). Juice Extraction

System: Mills and Diffusers - the Brazilian Experience. International

Sugar Journal, 190-202.

Paradis, P.-L., Rousse, D. R., Halle, S., Lamarche, L., & Quesada, G. (2015).

Thermal modeling of evacuated tube solar air collectors. Solar Enegy,

708-721.

Perry, R. H., Green, D. W., & Maloney, J. O. (1997). Pery's Chemical Engineer's

Handbook. New York: McGraw-Hill.

Petchers, N. (2012). Combined Hating, Cooling & Power Handbook:

Technologies & Applications . Routledge Chapman Hall.

Peterseim, J. H., Hellwig, U., Tadros, A., & White, S. (2013). Hybridisation

optimization of concentrating solar thermal and biomass power generation

facilities. Solar Energy, 203-214.

Peterseim, J. H., Herr, A., Miller, S., White, S., & O'Connell, S. A. (2014).

Hybridisation optimization of concentrating solar thermal and biomass

power generation facilities. Energy, 1-14.


104

Rayaprolu , K. (2009). Boiler for Power and Process . CRC Press.

Reid, M. J., & Rein, P. W. (1983). Steam Balance for the New Felixton II Mill.

Proceedings of the South African Sugar Technologists’ Association, (pp.

85-91). Durban.

Rein, P. W. (2007). Cane Sugar Engineering. Berlin: Bartens.

Sabiha, M. A., Saidur, R., Mekhilef, S., & Mahian, O. (2015). Progress and latest

developments of evacuated tube solar collectors. Renewable and

Sustainable Energy Reviews, 1038-1054.

SAIPPA. (2017). News. Retrieved November 6, 2017, from South African

Idependent Power Producers Association:

http://www.saippa.org.za/News.aspx

Schott. (2017). www.schott.com/borofloat. Retrieved August 12, 2017, from

www.schott.com:

http://www.schott.com/borofloat/english/attribute/optical/index.html

Siems, T. (2017). Airwasol Brochure, e-mail to W. Krog. [16 May 2017].

Simprotek. (2006). Drying Softare, Simprosys 3.0.

http://www.simprotek.com/Products.aspx.

Singels, A., McFarlane , S. A., Nicholson , R., Way, M., & Sithole , P. (2017).

Review of South African Sugarcane Production in the 2016/2017 season:

Light at the End of the Tunnel? . Durban: South African Sugar

Technologists’ Association Congress .

Smith, G. T., Davis, S. B., Madho, S., & Achary, M. (2015). Ninetieth Annual

Review of the Milling Season in Southern Africa (2014/2015). South

African Sugar Technologists’ Association Congress, (pp. 23-54). Durban.

Smith, G. T., Davis, S. B., Madho, S., & Achary, M. (2016). Ninety-First Annual

Review of the Milling Season in Southern Africa (2015/2016).

Proceedings of the South African Sugar Technologists' Association

Congress , (pp. 21-51). Durban .

Smithers, J. (2014). Review of sugarcane trash recovery systems for energy

cogeneration in South Africa. Renewable and Sustainable Energy Reviews,

915-925.

Solar Advice . (2016, June 22). Flat-Plate Solar Water Heating Collectors.

Retrieved October 16, 2017, from Solar Advice:

https://solaradvice.co.za/flat-plate-solar-water-heating

Solar Reserve. (n.d.). Photos. Retrieved October 15, 2017, from Solar Reserve:

http://www.solarreserve.com/en/newsroom/photos/crescent-dunes-solar-

energy-plant


105

Solargis. (2017). Duban Solar Resource Overview .

Sosa-Arnao, J. H., & Nebra, S. A. (2009). Bagasse Dryer Role in the Energy

Recovery of Water Tube Boilers. Drying Technology, 587-594.

Sosa-Arnoa, J. H., Correa, J. L., Silva, M. A., & Nebra, S. A. (2006). Sugar Cane

Bagasse Drying - a Review. International Sugar Journal.

Starzak , M., & Zizhou, N. (2015). Biorefinery Techno-Economic Modelling:

Sugar Mill and Ethanol Distillery Process Model. Technical Report No.

2210. Durban: Sugar Milling Research Institute .

Starzak, M., & Davis, S. B. (2016). MATLAB® modelling of a sugar mill: model

development and validation. South African Sugar Technologists'

Association, (p. 20). Durban.

Stine, W. B., & Geyer, M. (2001). Power from the Sun. John Wiley and Sons:

New Jersey.

Strumillo, C., & Kudra, T. (1986). Drying: Principles, Applications and Design.

Montreaux, Switzerland: Gordon and Breach Science Publishers.

Sun & Wind Energy. (n.d.). News: New parabolic trough collector from 3M and

Gossamer. Retrieved October 16, 2017, from Sun & Wind Energy: The

Platform for Renewable Energies :

http://www.sunwindenergy.com/news/new-parabolic-trough-collector-3m-

and-gossamer

Tawfik, A. A., El Rahman, A. K., Bayoumi, M. R., & Morsy, M. G. (2003).

Bagasse Drying: Advantages and Merits. Arab Region and Africa in the

World Sugar Context. Arwan.

Verein Detuscher Ingenieure. (2004). Solar heating for domestic water - General

Principles, system technology and use in residential building. Düsseldorf:

Verein Detuscher Ingenieure.

Vijayaraj, B., Saravanan, R., & Renganarayanan, S. (2007). Studies on thin layer

drying of bagasse. International Journal of Energy Research , 422-437.

Wang, Z., Yang, W., Qiu, F., Zhang, X., & Zhao, X. (2015). Solar water heating:

From theory, application, marketing and research. Renewable and

Sustainable Energy Reviews, 68-84.

Weiss, W., & Rommel, M. (2008). Process Heat Collectors . Gleisdorf : AEE

Intec .

Welz, C. (2017, March 9).

https://www.ise.fraunhofer.de/de/servicebereiche/testlab-solar-thermal-

systems.html#tabpanel-5. Retrieved April 3, 2017, from

https://www.ise.fraunhofer.de: https://www.ise.fraunhofer.de/en/press-


106

media/news/2017/fraunhofer-ise-develops-freeware-aircow-to-calculate-

solar-air-heating-collector-fields.html

Wienese, A. (2001). Boilers, Boiler Fuel and Boiler Efficiency . South African

Sugar Technologists' Association, (p. 7). Durban .

Zabaniotou, A. A. (2000). Simulation of Forestry Biomass Drying in a Rotary

Dryer. Drying Technology, 1415-1431.

Zoschke, T., Rohani, S., Fluri, T., Hu, Q., & Fang, Q. (2017). Parabolic Trough

Plant Performance in China with Focus on Comparison of Heat Transfer

Fluids HELISOL®5A and Therminol® VP-1. SolarPACES. Santiago .


Date post:	11-Sep-2021
Category:	Documents
Upload:	others
View:	5 times
Download:	0 times

Solar Live Steam Generation and Solar Bagasse Drying for South African Sugar Mills

Documents